π The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
TypeScript agentic-aiagentic-frameworkagentic-workflowagentsai-agentsai-assistantai-skillsautonomous-agentsclaudecodexdsh-pluginharnessmcp-servermulti-agentmulti-agent-systemsnpmskillsswarmswarm-intelligencetypescript
feat(plugins): cut per-turn footprint β user-invoked operational skills, description ceiling, footprint report (ADR-507) (#4588)
rUv828a8d78d97f
diff --git a/plugins/README.md b/plugins/README.mdindex 587aa72524..e4e3a6709a 100644--- a/plugins/README.md+++ b/plugins/README.md@@ -152,6 +152,24 @@ node scripts/audit-tool-descriptions.mjsCombined with [`verification/`](../verification/) (Ed25519-signed witness manifest, 103+ documented fixes attested), the plugin surface is regression-protected at three layers: install smoke (`npm i`), behavioral smoke (paired-tool round-trips), and presence attestation (every load-bearing line of every documented fix). See [`verification/README.md`](../verification/README.md) for the full stack.+## Per-turn footprint: user-invoked skills and short descriptions++Claude Code keeps the name and description of every skill, command and agent the model can invoke in context on every turn, used or not ([ADR-507](../v3/docs/adr/ADR-507-plugin-footprint-skills-commands.md)). Measure it with `node scripts/measure-plugin-footprint.mjs` (`--write` saves a report under `v3/docs/validation/`).++**Which skills are user-invoked.** A skill gets `disable-model-invocation: true` in its frontmatter when its job is something a person asks for by name, not something the model should reach for by itself. It still runs as `/<plugin>:<skill>`; it just leaves the model's listing. The kinds that qualify:++- release, publish and submission flows;+- dashboards, status pages and reports;+- benchmark runners and spend-gated learning or evolution runs;+- one-shot setup and install skills;+- sub-reports that a plugin command already reaches (for example the cost-tracker reports under `/ruflo-cost`).++A skill stays model-invoked when the model must find it on its own: memory search, swarm init guidance, safety scans, a gate another agent calls before it starts work (such as `cost-budget-check`), or anything an agent, command or hook in this repository tells the model to invoke. Before marking a skill, grep `plugins/*/agents`, `plugins/*/commands`, `plugins/*/skills` and `plugins/*/hooks` for its name: a model blocked by `disable-model-invocation` is told not to do the skill's work any other way.++Every user-invoked skill is listed under a "User-invoked skills" heading in its plugin README.++**Description ceiling.** A skill or command description is at most 300 characters: one sentence of what it does and one of when to use it ("Use when β¦"). `node scripts/audit-skill-frontmatter.mjs` (CI: all-plugins-smoke) enforces both rules.+## LicenseMITdiff --git a/plugins/ruflo-agntcy/.claude-plugin/plugin.json b/plugins/ruflo-agntcy/.claude-plugin/plugin.jsonindex cb31facb40..432207b6df 100644--- a/plugins/ruflo-agntcy/.claude-plugin/plugin.json+++ b/plugins/ruflo-agntcy/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-agntcy","description": "AGNTCY/Outshift Internet-of-Cognition runtime integration for Ruflo β SLIM secure transport for cross-host agent coordination, CASA intent-scoped tool authorization enforcement, optional IOC Layer 9 coordination events, and AGNTCY identity/observability spans (ADR-380). Optional, removable augmentation per ADR-150; degrades gracefully when AGNTCY/SLIM packages are not installed.",- "version": "0.2.5",+ "version": "0.2.6","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-agntcy/CHANGELOG.md b/plugins/ruflo-agntcy/CHANGELOG.mdindex 220be60d89..11ff7f04f4 100644--- a/plugins/ruflo-agntcy/CHANGELOG.md+++ b/plugins/ruflo-agntcy/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.2.6 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.2.5 β 2026-10-12- fix: the shared text screen now catches `.env` and e-mailed credentials, no longer flags a model's token talk (`--max-tokens`, `/oauth/token`, "token usage"), and no longer flags values that only name where a secret lives (`./service-account.json`, `--set-secrets=NAME=NAME:latest`); every rule stays linear (#4094, thanks @proffesor-for-testing)diff --git a/plugins/ruflo-agntcy/README.md b/plugins/ruflo-agntcy/README.mdindex dad2f53f89..474e5bc53b 100644--- a/plugins/ruflo-agntcy/README.md+++ b/plugins/ruflo-agntcy/README.md@@ -28,6 +28,14 @@ The remaining honest gaps: AGNTCY Identity has no JS/TS SDK (genuinely Go-only,| **IOC Layer 9 coordination events** | Β§4 | Optional semantic coordination events (Semantic Information Exchange, Cognition and Interoperability, Semantic Alignment Broadcast, Team Formation via Polling) layered on top of β never replacing β Ruflo's own `hive-mind_broadcast` / `hive-mind_consensus` / `coordination_consensus` orchestration. || **AGNTCY identity/observability** | Β§5 | OTel span attributes (`agent.identity`, `agent.capability`, `agent.intent`, `agent.parent`, `coordination.episode`, `authorization.decision`, `model.route`, `memory.provenance`, `evaluation.score`, `receipt.hash`) wired through the existing `ruflo-observability` plugin rather than a second tracing pipeline. Ruflo owns the two runtime-only attributes (`coordination.episode`, `authorization.decision`); the rest are emitted build-time by companion metaharness ADR-240. |+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. A status dashboard.++| Skill | Run it with |+|---|---|+| agntcy-status | `/ruflo-agntcy:agntcy-status` |+## Companion ADRThis plugin's runtime half is paired with metaharness repo ADR-240 (build-time half β AGNTCY identity generation, OASF export, semantic observability at manifest time). Neither ADR is complete without the other; see [ADR-380's "Companion" note](docs/adrs/ADR-380-agntcy-outshift-runtime-integration.md).diff --git a/plugins/ruflo-agntcy/scripts/smoke.sh b/plugins/ruflo-agntcy/scripts/smoke.shindex 7784884be1..c07e48a830 100755--- a/plugins/ruflo-agntcy/scripts/smoke.sh+++ b/plugins/ruflo-agntcy/scripts/smoke.sh@@ -8,9 +8,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "0. plugin.json declares 0.2.5"+step "0. plugin.json declares 0.2.6"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-[[ "$v" == "0.2.5" ]] && ok || bad "expected 0.2.5, got '$v'"+[[ "$v" == "0.2.6" ]] && ok || bad "expected 0.2.6, got '$v'"# M1. The mod (ADR-445 pattern): hooks module registered, files within the 500-line rulestep "M1. mod: hooks.json names register.ts, every hook file is present and under 500 lines"diff --git a/plugins/ruflo-agntcy/skills/agntcy-status/SKILL.md b/plugins/ruflo-agntcy/skills/agntcy-status/SKILL.mdindex 24123d3e33..41630be7e5 100644--- a/plugins/ruflo-agntcy/skills/agntcy-status/SKILL.md+++ b/plugins/ruflo-agntcy/skills/agntcy-status/SKILL.md@@ -1,8 +1,9 @@---name: agntcy-status-description: Show AGNTCY/SLIM/CASA integration status β whether the optional SLIM runtime package is installed, whether a SLIM endpoint is configured, which transport is the default, and what the agntcy mod has seen this session. Use when the user asks "is AGNTCY configured?", "show SLIM/CASA status", or "is AGNTCY/IOC integration active?". Read-only; makes no network call.+description: "Show AGNTCY/SLIM/CASA integration status: whether the SLIM runtime is installed, an endpoint configured, which transport is default, and what the agntcy mod saw this session. Use when the user asks whether AGNTCY is configured. Read-only; no network call."allowed-tools: Read Bashargument-hint: ""+disable-model-invocation: true---Report what can be measured on this machine, and say "unknown" for anything that cannot. Never invent a healthy status.diff --git a/plugins/ruflo-bbs-federation/.claude-plugin/plugin.json b/plugins/ruflo-bbs-federation/.claude-plugin/plugin.jsonindex 0c816e441b..8bd58d2e0f 100644--- a/plugins/ruflo-bbs-federation/.claude-plugin/plugin.json+++ b/plugins/ruflo-bbs-federation/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-bbs-federation","description": "AgentBBS cross-host agent federation for ruflo (ADR-164). Phase 1 exposes room coordination β federation_bbs_register / federation_bbs_publish / federation_bbs_watch / federation_bbs_human_join. Phase 2 (v3.40.0) adds signed cross-host federation β federation_bbs_identity (persistent Ed25519 node identity), federation_bbs_peer_add (pin a peer by nodeId + public key), federation_bbs_peers (audit/unpin), federation_bbs_serve (read-only pull endpoint), and federation_bbs_sync (pull pinned peers and union-merge what verifies). Every merged envelope is Ed25519-verified against a pinned key; unsigned/misattributed/oversize/over-hop envelopes are dropped and counted. Transport is HTTP pull and works on any network (Tailscale, LAN, VPN, loopback) β no tailnet required. Coordinates work ownership via claim messages (ClaimIssued/Released/Handoff/Ack). agentbbs lives in optionalDependencies; every handler gracefully degrades to {degraded:true} when the package is missing (ADR-150 pattern). As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.",- "version": "0.3.4",+ "version": "0.3.5","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-bbs-federation/CHANGELOG.md b/plugins/ruflo-bbs-federation/CHANGELOG.mdindex d2ee26a4b3..a717024ba1 100644--- a/plugins/ruflo-bbs-federation/CHANGELOG.md+++ b/plugins/ruflo-bbs-federation/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.3.5 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.3.4 β 2026-10-12- fix: the shared text screen now catches `.env` and e-mailed credentials, no longer flags a model's token talk (`--max-tokens`, `/oauth/token`, "token usage"), and no longer flags values that only name where a secret lives (`./service-account.json`, `--set-secrets=NAME=NAME:latest`); every rule stays linear (#4094, thanks @proffesor-for-testing)diff --git a/plugins/ruflo-bbs-federation/scripts/smoke.sh b/plugins/ruflo-bbs-federation/scripts/smoke.shindex f387a0242a..3f8b98588f 100755--- a/plugins/ruflo-bbs-federation/scripts/smoke.sh+++ b/plugins/ruflo-bbs-federation/scripts/smoke.sh@@ -1,5 +1,5 @@#!/usr/bin/env bash-# Structural smoke test for ruflo-bbs-federation v0.3.4: the plugin manifest and the mod (ADR-445 pattern).+# Structural smoke test for ruflo-bbs-federation v0.3.5: the plugin manifest and the mod (ADR-445 pattern).set -uROOT="$(cd "$(dirname "$0")/.." && pwd)"PASS=0@@ -8,10 +8,10 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares version 0.3.4 and the mod keywords"+step "1. plugin.json declares version 0.3.5 and the mod keywords"v=$(grep -E '"version"[[:space:]]*:' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.3.4" ]]; then- bad "expected 0.3.4, got '$v'"+if [[ "$v" != "0.3.5" ]]; then+ bad "expected 0.3.5, got '$v'"elsemiss=""for k in mod function-hooks guard; do grep -q "\"$k\"" "$ROOT/.claude-plugin/plugin.json" || miss="$miss $k"; donediff --git a/plugins/ruflo-bbs-federation/skills/cross-host-federation/SKILL.md b/plugins/ruflo-bbs-federation/skills/cross-host-federation/SKILL.mdindex 753d022a05..bf177f8faf 100644--- a/plugins/ruflo-bbs-federation/skills/cross-host-federation/SKILL.md+++ b/plugins/ruflo-bbs-federation/skills/cross-host-federation/SKILL.md@@ -1,10 +1,6 @@---name: cross-host-federation-description: >- Join and operate a signed cross-host agentbbs federation, and coordinate work claims across nodes.- Use when: connecting ruflo agents across machines, sharing status/tasks/results between hosts,- propagating work claims across a swarm, or standing up a federation hub.- Skip when: single-host local work with no other nodes to coordinate with.+description: Join and operate a signed cross-host agentbbs federation and coordinate work claims across nodes. Use when connecting ruflo agents across machines, sharing tasks or results between hosts, or standing up a federation hub; skip for single-host work.allowed-tools: mcp__plugin_ruflo-core_ruflo__federation_bbs_identity mcp__plugin_ruflo-core_ruflo__federation_bbs_peer_add mcp__plugin_ruflo-core_ruflo__federation_bbs_peers mcp__plugin_ruflo-core_ruflo__federation_bbs_serve mcp__plugin_ruflo-core_ruflo__federation_bbs_register mcp__plugin_ruflo-core_ruflo__federation_bbs_publish mcp__plugin_ruflo-core_ruflo__federation_bbs_sync mcp__plugin_ruflo-core_ruflo__federation_bbs_watch Readargument-hint: "[join|serve|status] [--hub <url>]"---diff --git a/plugins/ruflo-business-pods/.claude-plugin/plugin.json b/plugins/ruflo-business-pods/.claude-plugin/plugin.jsonindex f02abbba10..345e5ce508 100644--- a/plugins/ruflo-business-pods/.claude-plugin/plugin.json+++ b/plugins/ruflo-business-pods/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-business-pods","description": "ADR-164 Phase 2 β business-domain pod templates for the federated business autopilot. Ships the sales pod template (templates/sales.json) and a single-tick runner (scripts/pod-tick.mjs) that pre-flights templates via business_pod_validate, reserves budget through a file-based stub ledger (atomic SQLite tracker is Phase 3 per ADR-164.1), resolves each agent role against ruflo's agent registry, constructs dry-run prompts, and posts an envelope summary to the room via federation_bbs_publish. Dry-run by default β --live is reserved for Phase 3 Managed-Agent + claude-p wiring. As a mod (ADR-445 pattern): a tighten-only tool guard, a local slash command, and a status file the console shows.",- "version": "0.2.4",+ "version": "0.2.5","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-business-pods/CHANGELOG.md b/plugins/ruflo-business-pods/CHANGELOG.mdindex 2ca39581f7..c9ffed66bc 100644--- a/plugins/ruflo-business-pods/CHANGELOG.md+++ b/plugins/ruflo-business-pods/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.2.5 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.2.4 β 2026-10-12- fix: the shared text screen now catches `.env` and e-mailed credentials, no longer flags a model's token talk (`--max-tokens`, `/oauth/token`, "token usage"), and no longer flags values that only name where a secret lives (`./service-account.json`, `--set-secrets=NAME=NAME:latest`); every rule stays linear (#4094, thanks @proffesor-for-testing)diff --git a/plugins/ruflo-business-pods/scripts/smoke.sh b/plugins/ruflo-business-pods/scripts/smoke.shindex 0fe3ed4e6b..71676d35ec 100755--- a/plugins/ruflo-business-pods/scripts/smoke.sh+++ b/plugins/ruflo-business-pods/scripts/smoke.sh@@ -1,8 +1,8 @@#!/usr/bin/env bash-# Structural + functional smoke test for ruflo-business-pods v0.2.4+# Structural + functional smoke test for ruflo-business-pods v0.2.5# (ADR-164 Phase 2). Verifies:#-# 1. plugin.json declares 0.2.4 with the expected adr-164/phase-2 keywords+# 1. plugin.json declares 0.2.5 with the expected adr-164/phase-2 keywords# 2. templates/sales.json exists and validates against the pod-schema# 3. scripts/pod-tick.mjs exists, parses, and is executable# 4. Vitest suite passes (cli/__tests__/business-pod-tools.test.ts)@@ -23,10 +23,10 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.2.4 with adr-164 + phase-2 keywords"+step "1. plugin.json declares 0.2.5 with adr-164 + phase-2 keywords"v=$(grep -E '"version"' "$PLUGIN/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.2.4" ]]; then- bad "expected 0.2.4, got '$v'"+if [[ "$v" != "0.2.5" ]]; then+ bad "expected 0.2.5, got '$v'"elsemiss=""for k in ruflo pods business sales adr-164 adr-164.1 phase-2 pod-template pod-tick budget-reservation dry-run-default agent-registry-resolution; dodiff --git a/plugins/ruflo-business-pods/skills/pod-sales/SKILL.md b/plugins/ruflo-business-pods/skills/pod-sales/SKILL.mdindex 9645ed81cd..06d9c640c2 100644--- a/plugins/ruflo-business-pods/skills/pod-sales/SKILL.md+++ b/plugins/ruflo-business-pods/skills/pod-sales/SKILL.md@@ -1,6 +1,6 @@---name: pod-sales-description: Run one tick of the sales business-pod (ADR-164 Β§4.1, Phase 2). Loads templates/sales.json, validates it against the pod-schema, resolves agents against ruflo's agent registry, reserves budget via the Phase-2 file-based stub ledger (atomic SQLite tracker is Phase 3 per ADR-164.1), constructs per-agent dry-run prompts, posts a summary envelope to room "sales" via the federation_bbs_publish JSONL backing store, and emits a structured {podName, tickId, agentsRan, totalUsd, envelopeId, status} line for /loop ingestion. Dry-run by default; --live is reserved for Phase 3.+description: "Run one dry-run tick of the sales business-pod (ADR-164): validate templates/sales.json, reserve budget, post a summary envelope to the \"sales\" room and print a status line for /loop. Use when a person or a /loop asks for the next sales-pod tick."argument-hint: "[--pod-template <path>] [--base-path <dir>] [--dry-run|--live] [--budget-cap-usd <amt>] [--tick-id <id>]"allowed-tools: Bash---diff --git a/plugins/ruflo-core/.claude-plugin/plugin.json b/plugins/ruflo-core/.claude-plugin/plugin.jsonindex 26cb905e02..feca8c4a30 100644--- a/plugins/ruflo-core/.claude-plugin/plugin.json+++ b/plugins/ruflo-core/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-core","description": "Foundation plugin β registers the ruflo MCP server (300+ tools across memory/agentdb/embeddings/hooks/aidefence/neural/autopilot/browser/agent/swarm), provides 4 generalist agents (coder/researcher/reviewer/witness-curator), 5 skills (discover-plugins/init-project/ruflo-doctor/ruflo-status/witness), and a curated plugin-discovery catalog",- "version": "0.2.11",+ "version": "0.2.12","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-core/CHANGELOG.md b/plugins/ruflo-core/CHANGELOG.mdindex 403778e6fc..e20c1eddda 100644--- a/plugins/ruflo-core/CHANGELOG.md+++ b/plugins/ruflo-core/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.2.12 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.2.11 β 2026-10-14- perf: the MCP launcher runs the resolved local CLI inside its own process instead of spawning a second node (one process, about half the memory, a faster first answer: measured here with a stub CLI as 2 processes to 1, 89 MB to 46 MB, 30 ms to 19 ms spawn-to-initialize); `RUFLO_MCP_SPAWN=1` restores the child process, and the npx fallback still spawns (#3989, thanks @socrates6969). The real-CLI startup numbers in the PR (149 MB to 81 MB, 1.3 s to 0.7 s) were not re-measured.diff --git a/plugins/ruflo-core/README.md b/plugins/ruflo-core/README.mdindex e0f91d0336..d971fc37bc 100644--- a/plugins/ruflo-core/README.md+++ b/plugins/ruflo-core/README.md@@ -17,6 +17,15 @@ Foundation plugin. Registers the `ruflo` MCP server (314 tools), provides three- **Session Management**: Persistent sessions with cross-conversation learning- **Hooks**: PreToolUse / PostToolUse / PreCompact / Stop wired to claude-flow's auto-routing + learning loop. Defined at `plugins/ruflo-core/hooks/hooks.json` so the per-plugin loader picks them up on `/plugin install ruflo-core@ruflo` (per-plugin layout β fixes #1748 Issue 1; the marketplace-root copy at `.claude-plugin/hooks/hooks.json` is preserved for `claude --plugin-dir <repo-root>` users).+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. One-shot project setup and a status dashboard.++| Skill | Run it with |+|---|---|+| init-project | `/ruflo-core:init-project` |+| ruflo-status | `/ruflo-core:ruflo-status` |+## ConfigurationThe MCP server starts automatically when this plugin is active. Its launcher uses a built `@claude-flow/cli` from the project, an ancestor `node_modules`, or a global npm install reachable from `PATH` before falling back to `npx @latest`. Set `RUFLO_MCP_CLI_OVERRIDE` to a built `@claude-flow/cli/bin/cli.js` for a nonstandard install; an invalid override fails instead of running another version. Set `RUFLO_MCP_SKIP_NPX=1` to refuse the unpinned fallback. Override environment variables in `.mcp.json` as needed.diff --git a/plugins/ruflo-core/scripts/smoke.sh b/plugins/ruflo-core/scripts/smoke.shindex 15f3642a04..26523c38ab 100755--- a/plugins/ruflo-core/scripts/smoke.sh+++ b/plugins/ruflo-core/scripts/smoke.sh@@ -1,5 +1,5 @@#!/usr/bin/env bash-# Structural smoke test for ruflo-core v0.2.11 (ADR-0001).+# Structural smoke test for ruflo-core v0.2.12 (ADR-0001).set -uROOT="$(cd "$(dirname "$0")/.." && pwd)"PASS=0@@ -8,10 +8,10 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.2.11 with new keywords"+step "1. plugin.json declares 0.2.12 with new keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.2.11" ]]; then- bad "expected 0.2.11, got '$v'"+if [[ "$v" != "0.2.12" ]]; then+ bad "expected 0.2.12, got '$v'"elsemiss=""for k in foundation mcp-server plugin-catalog discovery; dodiff --git a/plugins/ruflo-core/skills/init-project/SKILL.md b/plugins/ruflo-core/skills/init-project/SKILL.mdindex 3d08300dc3..cc56ccfba2 100644--- a/plugins/ruflo-core/skills/init-project/SKILL.md+++ b/plugins/ruflo-core/skills/init-project/SKILL.md@@ -3,6 +3,7 @@ name: init-projectdescription: Initialize a new Ruflo project with MCP tools, hooks, and agent configuration. Use when setting up Ruflo in a fresh repo, or when the user says "init ruflo", "set up ruflo", or asks how to bootstrap the MCP server, hooks, and agent configs from scratch.argument-hint: "[--preset standard|minimal|full]"allowed-tools: Bash(npx *) Read Write Edit+disable-model-invocation: true---Run `npx @claude-flow/cli@latest init --wizard` to set up the project interactively, or `npx @claude-flow/cli@latest init --preset standard` for defaults.diff --git a/plugins/ruflo-core/skills/ruflo-status/SKILL.md b/plugins/ruflo-core/skills/ruflo-status/SKILL.mdindex 800e36d8cd..4f7c05c0b9 100644--- a/plugins/ruflo-core/skills/ruflo-status/SKILL.md+++ b/plugins/ruflo-core/skills/ruflo-status/SKILL.md@@ -3,6 +3,7 @@ name: ruflo-statusdescription: Diagnose Ruflo health, then report system, MCP server, and active-agent status without changing the installationargument-hint: "[--fix]"allowed-tools: Bash(npx *)+disable-model-invocation: true---# Ruflo statusdiff --git a/plugins/ruflo-core/skills/witness/SKILL.md b/plugins/ruflo-core/skills/witness/SKILL.mdindex 927cad63a9..9c848b3c36 100644--- a/plugins/ruflo-core/skills/witness/SKILL.md+++ b/plugins/ruflo-core/skills/witness/SKILL.md@@ -1,6 +1,6 @@---name: witness-description: Sign, verify, and track fix-marker regressions over time using an Ed25519 witness manifest. The default seed is public-commit-derived (integrity only); set RUFLO_WITNESS_SEED/RUFLO_WITNESS_PUBLIC_KEY for authorship. Works in any project β clone the toolkit, run init, register fixes, regen on each release.+description: Sign, verify and track fix-marker regressions over time with an Ed25519 witness manifest. Use when registering fixes or checking a release for regressed fixes. The default seed proves integrity only; set RUFLO_WITNESS_SEED/RUFLO_WITNESS_PUBLIC_KEY for authorship.argument-hint: "init|regen|verify|history [...]"allowed-tools: Bash(node *), Read, Write, Edit---diff --git a/plugins/ruflo-cost-tracker/.claude-plugin/plugin.json b/plugins/ruflo-cost-tracker/.claude-plugin/plugin.jsonindex 8909764e1f..848a480b9e 100644--- a/plugins/ruflo-cost-tracker/.claude-plugin/plugin.json+++ b/plugins/ruflo-cost-tracker/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-cost-tracker","description": "Token usage tracking, model cost attribution per agent, budget alerts, and optimization recommendations β uses memory_* (namespace-routed) for cost-tracking and cost-patterns; pairs with federation budget circuit breaker (ADR-097)",- "version": "0.29.0",+ "version": "0.30.0","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-cost-tracker/CHANGELOG.md b/plugins/ruflo-cost-tracker/CHANGELOG.mdindex 1e106accb2..89d932f4e9 100644--- a/plugins/ruflo-cost-tracker/CHANGELOG.md+++ b/plugins/ruflo-cost-tracker/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; older versions: `git log -- plugins/ruflo-cost-tracker`.+## 0.30.0 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.29.0 β 2026-10-29- feat: `openrouterApiKey` is a `userConfig` option with `sensitive: true` (ADR-506, #4587): entered masked, kept in the platform credential store, read through `readSecret` before the `OPENROUTER_API_KEY` environment fallback; never logged or put on a command line.diff --git a/plugins/ruflo-cost-tracker/README.md b/plugins/ruflo-cost-tracker/README.mdindex 2d0c35c823..423d2edf6c 100644--- a/plugins/ruflo-cost-tracker/README.md+++ b/plugins/ruflo-cost-tracker/README.md@@ -49,6 +49,27 @@ claude --plugin-dir plugins/ruflo-cost-trackerGrok prices are xAI list prices, not what a subscription bills; every Grok row is flagged **approximate** in the ledger output.+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. Each is also a `/ruflo-cost` subcommand (for example `/ruflo-cost anomaly`).++| Skill | Run it with |+|---|---|+| cost-anomaly | `/ruflo-cost-tracker:cost-anomaly` |+| cost-benchmark | `/ruflo-cost-tracker:cost-benchmark` |+| cost-burn | `/ruflo-cost-tracker:cost-burn` |+| cost-conversation | `/ruflo-cost-tracker:cost-conversation` |+| cost-counterfactual | `/ruflo-cost-tracker:cost-counterfactual` |+| cost-diff | `/ruflo-cost-tracker:cost-diff` |+| cost-export | `/ruflo-cost-tracker:cost-export` |+| cost-federation | `/ruflo-cost-tracker:cost-federation` |+| cost-health | `/ruflo-cost-tracker:cost-health` |+| cost-projection | `/ruflo-cost-tracker:cost-projection` |+| cost-report | `/ruflo-cost-tracker:cost-report` |+| cost-session | `/ruflo-cost-tracker:cost-session` |+| cost-summary | `/ruflo-cost-tracker:cost-summary` |+| cost-trend | `/ruflo-cost-tracker:cost-trend` |+## Commands (23 subcommands)```bashdiff --git a/plugins/ruflo-cost-tracker/agents/cost-analyst.md b/plugins/ruflo-cost-tracker/agents/cost-analyst.mdindex 0fa4dd5222..b1ee5409f4 100644--- a/plugins/ruflo-cost-tracker/agents/cost-analyst.md+++ b/plugins/ruflo-cost-tracker/agents/cost-analyst.md@@ -44,6 +44,8 @@ Model pricing per 1M tokens (Haiku/Sonnet/Opus Γ Input/Output/Cache-Write/Cache| `cost-booster-edit` | Apply a Tier 1 transform via `agent-booster.apply()` | When a transform is already classified as Tier 1 || `cost-compact-context` | Wrap `getTokenOptimizer().getCompactContext()` | Retrieval-augmented prompt compression |+The report skills (`cost-report`, `cost-conversation`, `cost-summary`, `cost-trend`, `cost-projection`, `cost-counterfactual`, `cost-burn`, `cost-anomaly`, `cost-health`, `cost-diff`, `cost-session`, `cost-export`, `cost-federation`, `cost-benchmark`) are user-invoked (ADR-507): the Skill tool refuses them. Tell the user which `/ruflo-cost` subcommand gives the report (for example `/ruflo-cost anomaly`) rather than producing it yourself.+`cost track` populates the namespace; everything else consumes it. Run `cost-track` first or other skills will operate on empty data.## Tools (MCP-routed primary path)diff --git a/plugins/ruflo-cost-tracker/scripts/smoke.sh b/plugins/ruflo-cost-tracker/scripts/smoke.shindex a873812cf0..6d13a4273a 100755--- a/plugins/ruflo-cost-tracker/scripts/smoke.sh+++ b/plugins/ruflo-cost-tracker/scripts/smoke.sh@@ -1,5 +1,5 @@#!/usr/bin/env bash-# Structural smoke test for ruflo-cost-tracker v0.29.0.+# Structural smoke test for ruflo-cost-tracker v0.30.0.set -u# `pwd -W` (Git Bash on Windows) gives C:/Users/...; plain pwd gives /c/Users/..., which a Windows node cannot open. Elsewhere -W fails and pwd answers.ROOT="$(cd "$(dirname "$0")/.." && { pwd -W 2>/dev/null || pwd; })"@@ -9,10 +9,10 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.29.0 with new keywords"+step "1. plugin.json declares 0.30.0 with new keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.29.0" ]]; then- bad "expected 0.29.0, got '$v'"+if [[ "$v" != "0.30.0" ]]; then+ bad "expected 0.30.0, got '$v'"elsemiss=""for k in namespace-routing mcp agentic-flow agent-booster tier1-routing model-routing benchmarking verified telemetry budget projection forecast counterfactual drift-detection trend-alert anomaly-detection outlier-detection health-check composite-gate auto-track stop-hook snapshot-diff pr-regression git-context traceability drill-down per-message; dodiff --git a/plugins/ruflo-cost-tracker/skills/cost-anomaly/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-anomaly/SKILL.mdindex b74f75ce1f..cb09143e50 100644--- a/plugins/ruflo-cost-tracker/skills/cost-anomaly/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-anomaly/SKILL.md@@ -1,8 +1,9 @@---name: cost-anomaly-description: MAD-based outlier detection on session spend. Robust to the very outliers it hunts (unlike mean+sigma). Surfaces specific anomalous sessions with modified-z scores; optional --alert-on-outliers exit code for CI gates. Distinct from cost-burn (aggregate trend) β this answers "which INDIVIDUAL session is the outlier?".+description: "MAD-based outlier detection on session spend: lists anomalous sessions with modified-z scores, with an optional --alert-on-outliers exit code for CI. Use when asked which individual session is the cost outlier; cost-burn covers the aggregate trend."argument-hint: "[--since 7d] [--threshold 3.5] [--alert-on-outliers N] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---Per-session outlier detection β the diagnostic counterpart to cost-burn'sdiff --git a/plugins/ruflo-cost-tracker/skills/cost-benchmark/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-benchmark/SKILL.mdindex cf7d592931..107c2e6f00 100644--- a/plugins/ruflo-cost-tracker/skills/cost-benchmark/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-benchmark/SKILL.md@@ -3,6 +3,7 @@ name: cost-benchmarkdescription: Run the corpus benchmark β booster locally, optional Gemini/Sonnet/Opus baselines β and persist a verifiable measured-vs-claimed tableargument-hint: "[--llm] [--anthropic]"allowed-tools: Bash+disable-model-invocation: true---# Cost Benchmarkdiff --git a/plugins/ruflo-cost-tracker/skills/cost-burn/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-burn/SKILL.mdindex fb2fcc0d9b..4b2e8879c1 100644--- a/plugins/ruflo-cost-tracker/skills/cost-burn/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-burn/SKILL.md@@ -3,6 +3,7 @@ name: cost-burndescription: Burn-rate trend over time with optional drift-alert exit code. Bins session spend into buckets, surfaces window-over-window delta, and can exit 1 when latest bucket exceeds prior mean by a configurable %. Distinct from `cost-trend` (benchmark drift); this tracks PRODUCTION spend trajectory.argument-hint: "[--bucket 1d] [--lookback 14d] [--alert-on-acceleration-pct 50] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---Burn-rate trend observability. The fourth leg of the cost-tracker forward-cost stack:diff --git a/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.mdindex d62fb8f7d2..34ffad0efa 100644--- a/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md@@ -3,6 +3,7 @@ name: cost-conversationdescription: Per-conversation cost view β list every session in cost-tracking with started-at, message count, top model, and total costargument-hint: ""allowed-tools: Bash+disable-model-invocation: true---# Cost per Conversationdiff --git a/plugins/ruflo-cost-tracker/skills/cost-counterfactual/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-counterfactual/SKILL.mdindex 669c5926e1..926f614e60 100644--- a/plugins/ruflo-cost-tracker/skills/cost-counterfactual/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-counterfactual/SKILL.md@@ -3,6 +3,7 @@ name: cost-counterfactualdescription: Multi-baseline counterfactual cost analysis. Compares actual session spend to hypothetical always-haiku / always-sonnet / always-opus routing baselines. Answers "is the routing earning its keep?" Negative savings flag over-escalation; positive savings quantify the router's win.argument-hint: "[--since 7d] [--baseline always-haiku|always-sonnet|always-opus|all] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---Multi-baseline counterfactual cost analysis. Pairs with the existing observability surface:diff --git a/plugins/ruflo-cost-tracker/skills/cost-diff/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-diff/SKILL.mdindex a036a66366..d954c60e89 100644--- a/plugins/ruflo-cost-tracker/skills/cost-diff/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-diff/SKILL.md@@ -3,6 +3,7 @@ name: cost-diffdescription: Snapshot delta between two cost-summary JSON outputs. PR-level cost regression detection β answers "what changed between these two specific snapshots?". Pairs with cost-summary's stable JSON contract.argument-hint: "--baseline <baseline.json> --current <current.json> [--alert-on-pct N] [--alert-on-usd N] [--alert-on-class-pct <class>:N[,<class>:N]] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---PR-level cost regression detection. Where cost-counterfactual comparesdiff --git a/plugins/ruflo-cost-tracker/skills/cost-export/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-export/SKILL.mdindex 15654a8a1f..3f70c325d9 100644--- a/plugins/ruflo-cost-tracker/skills/cost-export/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-export/SKILL.md@@ -3,6 +3,7 @@ name: cost-exportdescription: Export cost-tracking telemetry in Prometheus textfile or webhook JSON formats β for external observability (Grafana, Datadog, custom dashboards)argument-hint: "[--prometheus <path>] [--webhook <url>]"allowed-tools: Bash+disable-model-invocation: true---# Cost Exportdiff --git a/plugins/ruflo-cost-tracker/skills/cost-federation/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-federation/SKILL.mdindex f67f2c5c6b..4f92936a58 100644--- a/plugins/ruflo-cost-tracker/skills/cost-federation/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-federation/SKILL.md@@ -3,6 +3,7 @@ name: cost-federationdescription: Consumer-side wiring for ADR-097 Phase 3 federation_spend events β per-peer rolling windows + suspension-threshold checkargument-hint: ""allowed-tools: Bash+disable-model-invocation: true---# Cost Federationdiff --git a/plugins/ruflo-cost-tracker/skills/cost-health/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-health/SKILL.mdindex c7468619c1..ccc788a4f8 100644--- a/plugins/ruflo-cost-tracker/skills/cost-health/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-health/SKILL.md@@ -3,6 +3,7 @@ name: cost-healthdescription: Composite CI gate β runs cost-budget-check + cost-burn + cost-anomaly + cost-projection in parallel and surfaces a single combined health status with max exit code. The operationally-useful entry point β one shell-out covers all four alert ladders.argument-hint: "[--alert-acceleration 100] [--alert-outliers 1] [--alert-days-to-exhaust 14] [--skip burn,anomaly] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---The four CI-gate skills (budget / burn / anomaly / projection) each answerdiff --git a/plugins/ruflo-cost-tracker/skills/cost-projection/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-projection/SKILL.mdindex 9509d0acab..9960a70bea 100644--- a/plugins/ruflo-cost-tracker/skills/cost-projection/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-projection/SKILL.md@@ -3,6 +3,7 @@ name: cost-projectiondescription: Forward-looking spend extrapolation. Computes a USD-per-day rate from the recent measurement window, projects to 7d/30d/90d/365d horizons, and surfaces "days until budget exhausted" when a budget is configured. Predictive counterpart to `cost-budget-check` (reactive).argument-hint: "[--window 7d] [--horizons 7d,30d,90d,365d] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---Forward-looking cost projection that pairs with `cost-budget-check`:diff --git a/plugins/ruflo-cost-tracker/skills/cost-report/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-report/SKILL.mdindex 25f06bf3db..657ab29533 100644--- a/plugins/ruflo-cost-tracker/skills/cost-report/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-report/SKILL.md@@ -3,6 +3,7 @@ name: cost-reportdescription: Generate a cost report showing token usage and USD costs by agent and modelargument-hint: "[--period today]"allowed-tools: mcp__plugin_ruflo-core_ruflo__memory_search mcp__plugin_ruflo-core_ruflo__memory_list mcp__plugin_ruflo-core_ruflo__memory_retrieve mcp__plugin_ruflo-core_ruflo__agentdb_pattern-search mcp__plugin_ruflo-core_ruflo__agentdb_semantic-route Bash+disable-model-invocation: true---# Cost Reportdiff --git a/plugins/ruflo-cost-tracker/skills/cost-session/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-session/SKILL.mdindex bfa049645d..3553133335 100644--- a/plugins/ruflo-cost-tracker/skills/cost-session/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-session/SKILL.md@@ -3,6 +3,7 @@ name: cost-sessiondescription: Per-message cost breakdown within a single session. The drill-down companion to cost-anomaly β when an outlier session is flagged, this surfaces the specific expensive messages so operators can see whether the cost came from output tokens, cache writes, or model escalations.argument-hint: "[--session-id <id>] [--top 20] [--since <iso-ts>] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---When cost-anomaly flags a session as a >3.5Ο outlier, the next questiondiff --git a/plugins/ruflo-cost-tracker/skills/cost-summary/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-summary/SKILL.mdindex 77bef6ff89..7f2bfafe44 100644--- a/plugins/ruflo-cost-tracker/skills/cost-summary/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-summary/SKILL.md@@ -3,6 +3,7 @@ name: cost-summarydescription: Single-shot programmatic dump of all cost data β total spend, per-tier, top session, budget status, federation aggregate. JSON or markdown.argument-hint: "[--format json|markdown]"allowed-tools: Bash+disable-model-invocation: true---# Cost Summarydiff --git a/plugins/ruflo-cost-tracker/skills/cost-trend/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-trend/SKILL.mdindex c1fd56cade..c1ff1d23b3 100644--- a/plugins/ruflo-cost-tracker/skills/cost-trend/SKILL.md+++ b/plugins/ruflo-cost-tracker/skills/cost-trend/SKILL.md@@ -3,6 +3,7 @@ name: cost-trenddescription: Read every docs/benchmarks/runs/*.json and surface drift in win rate, latency, escalation rate, and LLM-baseline cost over timeargument-hint: ""allowed-tools: Bash+disable-model-invocation: true---# Cost Trenddiff --git a/plugins/ruflo-deepseek-harness/.claude-plugin/plugin.json b/plugins/ruflo-deepseek-harness/.claude-plugin/plugin.jsonindex b3c3e2b9d6..7737eef608 100644--- a/plugins/ruflo-deepseek-harness/.claude-plugin/plugin.json+++ b/plugins/ruflo-deepseek-harness/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-deepseek-harness","description": "DeepSeek model harness for ruflo β agents, commands, scripts, and skills for running and evaluating DeepSeek-backed workflows within the ruflo swarm.",- "version": "0.2.5",+ "version": "0.2.6","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-deepseek-harness/CHANGELOG.md b/plugins/ruflo-deepseek-harness/CHANGELOG.mdindex 211daaa08d..77f00434aa 100644--- a/plugins/ruflo-deepseek-harness/CHANGELOG.md+++ b/plugins/ruflo-deepseek-harness/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.2.6 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.2.5 β 2026-10-12- fix: the shared text screen now catches `.env` and e-mailed credentials, no longer flags a model's token talk (`--max-tokens`, `/oauth/token`, "token usage"), and no longer flags values that only name where a secret lives (`./service-account.json`, `--set-secrets=NAME=NAME:latest`); every rule stays linear (#4094, thanks @proffesor-for-testing)diff --git a/plugins/ruflo-deepseek-harness/scripts/smoke.sh b/plugins/ruflo-deepseek-harness/scripts/smoke.shindex f45c2b982c..416a847774 100755--- a/plugins/ruflo-deepseek-harness/scripts/smoke.sh+++ b/plugins/ruflo-deepseek-harness/scripts/smoke.sh@@ -9,8 +9,8 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares version 0.2.5"-grep -q '"version": "0.2.5"' "$ROOT/.claude-plugin/plugin.json" && ok || bad "version is not 0.2.5"+step "1. plugin.json declares version 0.2.6"+grep -q '"version": "0.2.6"' "$ROOT/.claude-plugin/plugin.json" && ok || bad "version is not 0.2.6"step "2. scripts, both skills, the command and the agent are present"miss=""diff --git a/plugins/ruflo-deepseek-harness/skills/deepseek-chat/SKILL.md b/plugins/ruflo-deepseek-harness/skills/deepseek-chat/SKILL.mdindex 0183caeaa4..b9913d304f 100644--- a/plugins/ruflo-deepseek-harness/skills/deepseek-chat/SKILL.md+++ b/plugins/ruflo-deepseek-harness/skills/deepseek-chat/SKILL.md@@ -1,6 +1,6 @@---name: deepseek-chat-description: "One-shot chat completion against DeepSeek's `deepseek-chat` model via the OpenAI-compatible /v1/chat/completions endpoint. Reads DEEPSEEK_API_KEY from the environment; degrades gracefully (exit 0 with a JSON status:degraded envelope) when the key is missing or the API is unreachable. Use for non-reasoning tasks β summarization, extraction, quick classification β where deepseek-reasoner would be overkill."+description: One-shot chat completion against DeepSeek's `deepseek-chat` model. Use for non-reasoning tasks such as summarization, extraction or quick classification. Reads DEEPSEEK_API_KEY; returns status:degraded when the key is missing or the API is unreachable.argument-hint: "--prompt <text> [--system <text>] [--model deepseek-chat] [--temperature 0.7] [--max-tokens 1024] [--format table|json] [--alert-on-error]"allowed-tools: Bash---diff --git a/plugins/ruflo-deepseek-harness/skills/deepseek-reason/SKILL.md b/plugins/ruflo-deepseek-harness/skills/deepseek-reason/SKILL.mdindex 9c64fa1d9b..71d8b80609 100644--- a/plugins/ruflo-deepseek-harness/skills/deepseek-reason/SKILL.md+++ b/plugins/ruflo-deepseek-harness/skills/deepseek-reason/SKILL.md@@ -1,6 +1,6 @@---name: deepseek-reason-description: "Reasoning-mode completion against DeepSeek's `deepseek-reasoner` model (R1) via /v1/chat/completions. Surfaces the model's chain-of-thought (`reasoning_content`) separately from the final answer (`content`), so callers can display or discard the CoT without re-parsing. Reads DEEPSEEK_API_KEY; degrades gracefully (exit 0 with status:degraded envelope) when unset or the API is unreachable. Ignores temperature/top_p per DeepSeek's spec for reasoner models."+description: Reasoning completion against DeepSeek's `deepseek-reasoner` (R1), returning the chain of thought apart from the answer. Use when a task needs step-by-step reasoning from DeepSeek. Reads DEEPSEEK_API_KEY; returns status:degraded when unset or unreachable.argument-hint: "--prompt <text> [--system <text>] [--model deepseek-reasoner] [--max-tokens 4096] [--show-reasoning] [--format table|json] [--alert-on-error]"allowed-tools: Bash---diff --git a/plugins/ruflo-federation/.claude-plugin/plugin.json b/plugins/ruflo-federation/.claude-plugin/plugin.jsonindex eea185e258..8297f08e32 100644--- a/plugins/ruflo-federation/.claude-plugin/plugin.json+++ b/plugins/ruflo-federation/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-federation","description": "Cross-installation agent federation with zero-trust security, peer discovery, consensus-based task routing, and per-call budget circuit breaker (ADR-097)",- "version": "0.3.6",+ "version": "0.3.7","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-federation/CHANGELOG.md b/plugins/ruflo-federation/CHANGELOG.mdindex e8403f9851..8500988323 100644--- a/plugins/ruflo-federation/CHANGELOG.md+++ b/plugins/ruflo-federation/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.3.7 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.3.6 β 2026-10-12- fix: the shared text screen now catches `.env` and e-mailed credentials, no longer flags a model's token talk (`--max-tokens`, `/oauth/token`, "token usage"), and no longer flags values that only name where a secret lives (`./service-account.json`, `--set-secrets=NAME=NAME:latest`); every rule stays linear (#4094, thanks @proffesor-for-testing)diff --git a/plugins/ruflo-federation/README.md b/plugins/ruflo-federation/README.mdindex 29a72f74b7..f349b9200a 100644--- a/plugins/ruflo-federation/README.md+++ b/plugins/ruflo-federation/README.md@@ -56,6 +56,14 @@ Phase 1 enforces at the **send** side. Phase 2 (peer state machine: ACTIVE / SUS| `federation-status` | "is federation healthy?", "show peers", "federation status" || `federation-audit` | Query audit logs with compliance / severity / date filters |+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. One-shot setup that generates this node's keypair.++| Skill | Run it with |+|---|---|+| federation-init | `/ruflo-federation:federation-init` |+## Agents| Agent | Description |diff --git a/plugins/ruflo-federation/scripts/smoke.sh b/plugins/ruflo-federation/scripts/smoke.shindex cab6181c9d..004ab05e92 100755--- a/plugins/ruflo-federation/scripts/smoke.sh+++ b/plugins/ruflo-federation/scripts/smoke.sh@@ -7,9 +7,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.3.6 with new keywords"+step "1. plugin.json declares 0.3.7 with new keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.3.6" ]]; then bad "expected 0.3.6, got '$v'"; else+if [[ "$v" != "0.3.7" ]]; then bad "expected 0.3.7, got '$v'"; elsemiss=""for k in mcp pii-pipeline audit-log; dogrep -q "\"$k\"" "$ROOT/.claude-plugin/plugin.json" || miss="$miss $k"diff --git a/plugins/ruflo-federation/skills/federation-init/SKILL.md b/plugins/ruflo-federation/skills/federation-init/SKILL.mdindex cac1abb3a6..e320fa706f 100644--- a/plugins/ruflo-federation/skills/federation-init/SKILL.md+++ b/plugins/ruflo-federation/skills/federation-init/SKILL.md@@ -3,6 +3,7 @@ name: federation-initdescription: Initialize federation on this node β generate keypair and configure peersallowed-tools: Bash(npx *) mcp__plugin_ruflo-core_ruflo__memory_store mcp__plugin_ruflo-core_ruflo__hooks_post-task Read Writeargument-hint: "[--compliance hipaa|soc2|gdpr|none]"+disable-model-invocation: true---Initialize this node for federation. Generates an ed25519 keypair, creates the federation config, and optionally sets a compliance mode.diff --git a/plugins/ruflo-metaharness/.claude-plugin/plugin.json b/plugins/ruflo-metaharness/.claude-plugin/plugin.jsonindex a48a2a3570..f44a29dc39 100644--- a/plugins/ruflo-metaharness/.claude-plugin/plugin.json+++ b/plugins/ruflo-metaharness/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-metaharness","description": "MetaHarness integration for ruflo β surfaces score/genome/mint/mcp-scan/threat-model via skills; pairs with @metaharness/router (ADR-148/149) for cost-optimal model routing; honors ADR-150's architectural constraint that MetaHarness remains an optional augmentation, never a required runtime dependency. As a mod (ADR-445): confirms flywheel promotion, keeps secrets out of inputs, /metaharness-mod, and a status file the console shows",- "version": "0.3.0",+ "version": "0.4.0","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-metaharness/CHANGELOG.md b/plugins/ruflo-metaharness/CHANGELOG.mdindex 7a865bb2a0..ba06edfe76 100644--- a/plugins/ruflo-metaharness/CHANGELOG.md+++ b/plugins/ruflo-metaharness/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.4.0 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.3.0 β 2026-10-29- feat: `openrouterApiKey` is a `userConfig` option with `sensitive: true` (ADR-506, #4587), read through `readSecret` before the `OPENROUTER_API_KEY` environment fallback.diff --git a/plugins/ruflo-metaharness/README.md b/plugins/ruflo-metaharness/README.mdindex 505c5ce73e..793996588a 100644--- a/plugins/ruflo-metaharness/README.md+++ b/plugins/ruflo-metaharness/README.md@@ -29,6 +29,19 @@ MetaHarness integration plugin for ruflo. Surfaces the upstream `metaharness` /| `harness-learn` | `/harness-learn --host <h> --model <m> --slice <manifest> [--repo <checkout>] [--run]` | [email protected] / upstream ADR-235 β GEPA learning run; $0 dry-run default, `--run` to spend; needs a metaharness repo checkout || `harness-gepa` | `/harness-gepa --op genome\|validate\|render\|analyze [--path <genome.json>]` | [email protected] GEPA library surface β genome load/validate/render + transcript failure analysis; `gepaOptimize` stays library-only |+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. They run benchmarks, spend-gated learning or evolution runs, scheduled audits, or a one-shot scaffold.++| Skill | Run it with |+|---|---|+| harness-bench | `/ruflo-metaharness:harness-bench` |+| harness-evolve | `/ruflo-metaharness:harness-evolve` |+| harness-learn | `/ruflo-metaharness:harness-learn` |+| harness-mint | `/ruflo-metaharness:harness-mint` |+| harness-oia-audit | `/ruflo-metaharness:harness-oia-audit` |+| harness-security-bench | `/ruflo-metaharness:harness-security-bench` |+## Phase-0 baseline (ruflo itself, 2026-06-16)```jsondiff --git a/plugins/ruflo-metaharness/agents/metaharness-architect.md b/plugins/ruflo-metaharness/agents/metaharness-architect.mdindex 60637ce311..e764e2c85e 100644--- a/plugins/ruflo-metaharness/agents/metaharness-architect.md+++ b/plugins/ruflo-metaharness/agents/metaharness-architect.md@@ -36,6 +36,8 @@ needs its own ADR.| `harness-oia-audit` | Composite weekly audit worker (iter 7) | Cron-scheduled; bundles oia+threat+mcp into one timestamped record in `metaharness-audit` namespace || `harness-mint` | Scaffold a custom harness | User wants to fork; ALWAYS dry-run first, never write to project root |+`harness-oia-audit` and `harness-mint` are user-invoked (ADR-507), as are `harness-bench`, `harness-evolve`, `harness-learn` and `harness-security-bench`: the Skill tool refuses them. Tell the user which `/ruflo-metaharness:<skill>` to run rather than running it yourself.+## Tools (subprocess-only, no library imports)- All skills shell out to the pinned `metaharness` / `harness` binaries (`metaharness@~0.4.1`, local install or one-time versioned cache β never `@latest`) via the `_harness.mjs` shared helper.diff --git a/plugins/ruflo-metaharness/scripts/smoke.sh b/plugins/ruflo-metaharness/scripts/smoke.shindex f730afa610..dc4f6ba015 100755--- a/plugins/ruflo-metaharness/scripts/smoke.sh+++ b/plugins/ruflo-metaharness/scripts/smoke.sh@@ -1,5 +1,5 @@#!/usr/bin/env bash-# Structural smoke test for ruflo-metaharness v0.3.0 (ADR-150 Phase 1).+# Structural smoke test for ruflo-metaharness v0.4.0 (ADR-150 Phase 1).set -uROOT="$(cd "$(dirname "$0")/.." && pwd)"PASS=0@@ -32,10 +32,10 @@ elsebad "extraction-regex-rot: EXPECTED_TOOLS=$EXPECTED_TOOLS EXPECTED_SUBS=$EXPECTED_SUBS"fi-step "1. plugin.json declares 0.3.0 with adr-150 keywords"+step "1. plugin.json declares 0.4.0 with adr-150 keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.3.0" ]]; then- bad "expected 0.3.0, got '$v'"+if [[ "$v" != "0.4.0" ]]; then+ bad "expected 0.4.0, got '$v'"elsemiss=""for k in ruflo metaharness harness scorecard genome mcp-scan threat-model router adr-150 adr-148 adr-149 optional-dependency graceful-degradation subprocess phase-1-mvp; do@@ -483,7 +483,7 @@ step "17z74a. plugin helper pins accept the CLI-declared ranges β no tool-timemiss=""# 17z74 compares package.json files with each other; the helpers' own# *_PIN_VERSION constants (what findLocalPackageDir accepts) were never-# compared with them and drifted (~0.3.0 / ~0.8.0 vs declared ^0.4.1 /+# compared with them and drifted (~0.4.0 / ~0.8.0 vs declared ^0.4.1 /# ~0.10.2), so every metaharness_* call npm-installed an older release next# to the one ruflo ships. Hermetic runtime gate: stub packages at the declared# versions, published layout, npm/npx trap.@@ -1287,7 +1287,7 @@ miss=""DISP="$ROOT/../../v3/@claude-flow/cli/src/commands/metaharness.ts"WRAPPER="$ROOT/../../v3/@claude-flow/cli/src/mcp-tools/metaharness-tools.ts"# All 13 dispatchable subcommands present (iter-73 had 10; @metaharness/redblue-# added redblue.mjs; [email protected]/[email protected] added learn.mjs + gepa.mjs).+# added redblue.mjs; [email protected]/[email protected] added learn.mjs + gepa.mjs).for sub in score genome mcp-scan threat-model oia-audit audit-list audit-trend similarity drift-from-history mint redblue learn gepa; dogrep -q "${sub}" "$DISP" 2>/dev/null || miss="$miss subcommand-${sub}-not-listed"done@@ -2199,7 +2199,7 @@ grep -q "result has 'degraded'" "$F" || miss="$miss no-degraded-assertion"grep -q "result has 'exitCode'" "$F" || miss="$miss no-exitcode-assertion"# All tool names enumerated (similarity iter 36, drift_from_history iter 54,# ADR-153 added bench/evolve/security_bench, @metaharness/redblue added redblue,-# [email protected]/[email protected] added learn + gepa, ADR-322 added flywheel)+# [email protected]/[email protected] added learn + gepa, ADR-322 added flywheel)for tool in metaharness_score metaharness_genome metaharness_mcp_scan metaharness_threat_model metaharness_oia_audit metaharness_audit_list metaharness_audit_trend metaharness_similarity metaharness_drift_from_history metaharness_bench metaharness_evolve metaharness_security_bench metaharness_redblue metaharness_learn metaharness_gepa metaharness_flywheel; dogrep -q "${tool}" "$F" || miss="$miss missing-${tool}"donediff --git a/plugins/ruflo-metaharness/skills/harness-bench/SKILL.md b/plugins/ruflo-metaharness/skills/harness-bench/SKILL.mdindex 35d716dfc8..c00c484d25 100644--- a/plugins/ruflo-metaharness/skills/harness-bench/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-bench/SKILL.md@@ -1,8 +1,9 @@---name: harness-bench-description: Manage `@metaharness/darwin` bench suites β `bench create {repo}` scaffolds a JSON suite from a repo's test corpus; `bench verify {suite.json}` checks suite well-formedness. Bench suites are the fixed evaluation corpora that `harness-evolve --bench {suite.json}` scores variants against, decoupling evolution from the repo's natural tests. Degrades gracefully when @metaharness/darwin is absent.+description: Create (`bench create {repo}`) or verify (`bench verify {suite.json}`) a @metaharness/darwin bench suite, the fixed corpus `harness-evolve --bench` scores variants against. Use when preparing an evolution benchmark. Degrades gracefully without darwin.argument-hint: "--op create --repo <path> [--out <path>] | --op verify --suite <path>"allowed-tools: Bash+disable-model-invocation: true---Surfaces `metaharness-darwin bench <create|verify>` β the supporting verbdiff --git a/plugins/ruflo-metaharness/skills/harness-drift-from-history/SKILL.md b/plugins/ruflo-metaharness/skills/harness-drift-from-history/SKILL.mdindex 5e48feb907..f51900db9c 100644--- a/plugins/ruflo-metaharness/skills/harness-drift-from-history/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-drift-from-history/SKILL.md@@ -1,6 +1,6 @@---name: harness-drift-from-history-description: One-command drift detection. Composes audit-list + oia-audit + audit-trend into a single primitive β finds the most recent audit in `metaharness-audit` namespace, runs a fresh audit against the current repo, diffs them via ADR-152 Β§3.1 similarity, and alerts when structural distance crosses `--threshold`. Iter 53 of ADR-150 deep integration.+description: "One-command drift detection: run a fresh metaharness audit, diff it against the latest stored audit by ADR-152 similarity, and alert when the distance crosses `--threshold`. Use when checking whether a harness has drifted since its last audit."argument-hint: "[--path .] [--baseline-since 7d] [--threshold 0.95] [--dry-run] [--format json|table]"allowed-tools: Bash---diff --git a/plugins/ruflo-metaharness/skills/harness-evolve/SKILL.md b/plugins/ruflo-metaharness/skills/harness-evolve/SKILL.mdindex 96d4260b0b..224469780f 100644--- a/plugins/ruflo-metaharness/skills/harness-evolve/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-evolve/SKILL.md@@ -1,8 +1,9 @@---name: harness-evolve-description: Run `@metaharness/darwin evolve {repo}` to mutate a harness's seven policy surfaces (planner/contextBuilder/reviewer/retryPolicy/toolPolicy/memoryPolicy/scorePolicy), sandbox-score each variant, and promote only measured wins. The model is frozen; the harness evolves. Closes the loop ADR-150 opens (score+genome describe; evolve changes). Degrades gracefully when @metaharness/darwin is absent (ADR-150 + ADR-153 architectural constraints).+description: "Run `@metaharness/darwin evolve {repo}`: mutate a harness's seven policy surfaces, sandbox-score each variant and promote only measured wins. Use when evolving a harness against its tests or a bench suite. Degrades gracefully without darwin (ADR-150/153)."argument-hint: "--repo <path> [--generations 3] [--children 3] [--concurrency 2] [--sandbox real|mock|agent] [--selection pareto|quality-diversity|...] [--mutator deterministic|ruvllm] [--diagnose] [--confirm]"allowed-tools: Bash+disable-model-invocation: true---Surfaces the upstream `metaharness-darwin evolve` CLI as a ruflo skill. Thediff --git a/plugins/ruflo-metaharness/skills/harness-gepa/SKILL.md b/plugins/ruflo-metaharness/skills/harness-gepa/SKILL.mdindex a930fcb50c..fc83dad430 100644--- a/plugins/ruflo-metaharness/skills/harness-gepa/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-gepa/SKILL.md@@ -1,6 +1,6 @@---name: harness-gepa-description: "Inspect and audit GEPA genomes via the `@metaharness/darwin/gepa` library entry (darwin 0.8.0) β load/validate a genome (default is the shipped cand-6 promotion), render the system prompt a genome compiles to, or classify failure modes in a run transcript. The `gepaOptimize` loop itself is library-only (bring your own evaluator) and not surfaced here β use `harness-evolve` for sandbox-scored evolution. Degrades gracefully when @metaharness/darwin is absent."+description: Load, validate or render a GEPA genome, or classify failure modes in a run transcript, via @metaharness/darwin/gepa. Use when inspecting or auditing a genome; use harness-evolve for sandbox-scored evolution. Degrades gracefully without darwin.argument-hint: "--op genome|validate|render|analyze [--path <genome.json>] [--transcript <t.json>] [--alert-on-invalid]"allowed-tools: Bash---diff --git a/plugins/ruflo-metaharness/skills/harness-learn/SKILL.md b/plugins/ruflo-metaharness/skills/harness-learn/SKILL.mdindex 46fed06da3..488f281176 100644--- a/plugins/ruflo-metaharness/skills/harness-learn/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-learn/SKILL.md@@ -1,8 +1,9 @@---name: harness-learn-description: Run a GEPA learning cycle via `metaharness learn` (upstream ADR-235, [email protected]) β optimizes a harness genome against a SWE-bench-style slice manifest. $0 dry-run by default; `--run` is the explicit spend opt-in. Requires a metaharness repo checkout (`--repo` or $METAHARNESS_REPO) β without one it reports `checkout-required` with clone instructions. Degrades gracefully when metaharness is absent.+description: Run a GEPA learning cycle via `metaharness learn`, optimizing a harness genome against a SWE-bench-style slice. Use when training a harness genome. $0 dry-run by default; `--run` opts into spend. Needs a metaharness checkout (--repo or $METAHARNESS_REPO).argument-hint: "--host <h> --model <m> --slice <manifest> [--repo <checkout>] [--run] [--alert-on-fail]"allowed-tools: Bash+disable-model-invocation: true---Surfaces `metaharness learn` β the upstream GEPA learning harness thatdiff --git a/plugins/ruflo-metaharness/skills/harness-mint/SKILL.md b/plugins/ruflo-metaharness/skills/harness-mint/SKILL.mdindex b166b05718..53ed3ca068 100644--- a/plugins/ruflo-metaharness/skills/harness-mint/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-mint/SKILL.md@@ -1,8 +1,9 @@---name: harness-mint-description: Scaffold a custom AI agent harness via `metaharness new {name} --template {id} --host {id}`. Defaults to DRY-RUN (no writes) unless --confirm is passed. Refuses to write to the calling repo root or anywhere inside it. Honors ADR-150 architectural constraint + ruflo's "destructive-action confirmation" pattern.+description: Scaffold a custom AI agent harness via `metaharness new {name} --template {id} --host {id}`. Use when the user wants a new standalone harness. Dry-run unless --confirm is passed, and refuses to write inside the calling repo.argument-hint: "--name <id> --template <vertical:coding|minimal|β¦> [--host claude-code|codex|β¦] [--target /abs/path] [--confirm] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---The one write-capable skill in the plugin. Every other skill isdiff --git a/plugins/ruflo-metaharness/skills/harness-oia-audit/SKILL.md b/plugins/ruflo-metaharness/skills/harness-oia-audit/SKILL.mdindex 7152debec4..9d6023af7b 100644--- a/plugins/ruflo-metaharness/skills/harness-oia-audit/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-oia-audit/SKILL.md@@ -3,6 +3,7 @@ name: harness-oia-auditdescription: Composite Phase-2 audit worker (ADR-150). Bundles harness oia-manifest + threat-model + mcp-scan into one timestamped audit record stored in the `metaharness-audit` memory namespace. Designed for cron-scheduled drift detection.argument-hint: "[--path .] [--dry-run] [--alert-on-worst clean|low|medium|high] [--format table|json]"allowed-tools: Bash+disable-model-invocation: true---The 13th worker (ADR-150 Phase 2) β runs three MetaHarness staticdiff --git a/plugins/ruflo-metaharness/skills/harness-score/SKILL.md b/plugins/ruflo-metaharness/skills/harness-score/SKILL.mdindex 9c0f93dabd..e97d65fd94 100644--- a/plugins/ruflo-metaharness/skills/harness-score/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-score/SKILL.md@@ -1,6 +1,6 @@---name: harness-score-description: 5-dimension harness readiness scorecard from `metaharness score {path}`. Returns harnessFit / compileConfidence / taskCoverage / toolSafety / memoryUsefulness + estCostPerRunUsd + scaffoldReady. Pure-read; subprocess invocation; degrades gracefully when MetaHarness is absent (ADR-150 architectural constraint).+description: 5-dimension harness readiness scorecard from `metaharness score {path}` (harnessFit, compileConfidence, taskCoverage, toolSafety, memoryUsefulness, plus cost per run). Use when rating a repo as an agent harness. Pure-read; degrades gracefully without MetaHarness.argument-hint: "[--path .] [--alert-on-fit-below 70] [--format table|json]"allowed-tools: Bash---diff --git a/plugins/ruflo-metaharness/skills/harness-security-bench/SKILL.md b/plugins/ruflo-metaharness/skills/harness-security-bench/SKILL.mdindex 9a19e1d9f0..45d74693c0 100644--- a/plugins/ruflo-metaharness/skills/harness-security-bench/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-security-bench/SKILL.md@@ -1,8 +1,9 @@---name: harness-security-bench-description: Run `@metaharness/darwin security bench` (upstream "Darwin Shield" / ADR-155) β evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on TPR/FPR/patch-pass/repro/unsafe vs four baselines (B0 static, B1 LLM-single-pass, B2 fixed-agent, B3 Darwin-champion). Closest reference implementation for ruflo's own ADR-155 nightly self-learning security harness (PR #2417). Degrades gracefully when @metaharness/darwin is absent.+description: "Run `@metaharness/darwin security bench` (Darwin Shield): evolve a security-detection harness on a 10-vuln/9-decoy corpus and grade TPR/FPR/patch/repro against four baselines. Use when benchmarking security-harness detection. Degrades gracefully without darwin."argument-hint: "[--population 2] [--cycles 1] [--seed N] [--alert-on-fail]"allowed-tools: Bash+disable-model-invocation: true---Surfaces the upstream `metaharness-darwin security bench` command. **This isdiff --git a/plugins/ruflo-metaharness/skills/harness-similarity/SKILL.md b/plugins/ruflo-metaharness/skills/harness-similarity/SKILL.mdindex 29eb8ea8f6..7e098895be 100644--- a/plugins/ruflo-metaharness/skills/harness-similarity/SKILL.md+++ b/plugins/ruflo-metaharness/skills/harness-similarity/SKILL.md@@ -1,6 +1,6 @@---name: harness-similarity-description: ADR-152 β weighted similarity between two harness fingerprints (genome + score JSON). Returns overall score in [0,1] plus per-component breakdown (cosine over 9 numerics, categorical agreement over 4 enums, jaccard over agent_topology). Unblocks ADR-151 Β§3.2 Recommender, Β§3.3 Drift Detection, Β§3.5 Plugin Compat. Pure-TS, no `@metaharness/*` dep β preserves ADR-150's four architectural constraints.+description: "Weighted similarity (ADR-152) between two harness fingerprints (genome + score JSON): an overall score in [0,1] plus per-component breakdown. Use when comparing two harnesses or detecting drift between them. Pure TypeScript; needs no @metaharness package."argument-hint: "(--a a.json --b b.json | --a-key X --b-key Y) [--per-dimension] [--alert-below 0.5] [--format json|table]"allowed-tools: Bash---diff --git a/plugins/ruflo-plugin-creator/.claude-plugin/plugin.json b/plugins/ruflo-plugin-creator/.claude-plugin/plugin.jsonindex 24795128c1..8f4a704b2a 100644--- a/plugins/ruflo-plugin-creator/.claude-plugin/plugin.json+++ b/plugins/ruflo-plugin-creator/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-plugin-creator","description": "Scaffold, validate, and publish new Claude Code plugins with the canonical plugin contract β now including governed Claude Code mods (function hooks), ADR + smoke + Compatibility + namespace coordination + MCP-tool drift warnings. As a mod (ADR-445): /creator-mod (status, reserved names, a plugin contract check) and a status file the console shows",- "version": "0.4.3",+ "version": "0.4.4","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-plugin-creator/CHANGELOG.md b/plugins/ruflo-plugin-creator/CHANGELOG.mdindex 832ed5dbc6..292784a7cc 100644--- a/plugins/ruflo-plugin-creator/CHANGELOG.md+++ b/plugins/ruflo-plugin-creator/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.4.4 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.4.3 β 2026-10-11- fix: create-plugin required-fields list now matches `audit-plugin-manifest` (adds `keywords`, name equals directory, bump the version on every change) and the agentdb tool count is 20- fix: validate-plugin check 10 now verifies referenced tools exist (`mcp tools`), not only that they carry the right prefixdiff --git a/plugins/ruflo-plugin-creator/scripts/smoke.sh b/plugins/ruflo-plugin-creator/scripts/smoke.shindex 038d893730..a943d46bd7 100755--- a/plugins/ruflo-plugin-creator/scripts/smoke.sh+++ b/plugins/ruflo-plugin-creator/scripts/smoke.sh@@ -7,9 +7,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.4.3 with new keywords"+step "1. plugin.json declares 0.4.4 with new keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.4.3" ]]; then bad "expected 0.4.3, got '$v'"; else+if [[ "$v" != "0.4.4" ]]; then bad "expected 0.4.4, got '$v'"; elsemiss=""for k in mcp scaffolding contract-bootstrap; dogrep -q "\"$k\"" "$ROOT/.claude-plugin/plugin.json" || miss="$miss $k"diff --git a/plugins/ruflo-plugin-creator/skills/create-mod/SKILL.md b/plugins/ruflo-plugin-creator/skills/create-mod/SKILL.mdindex 29101de781..dd72f6d531 100644--- a/plugins/ruflo-plugin-creator/skills/create-mod/SKILL.md+++ b/plugins/ruflo-plugin-creator/skills/create-mod/SKILL.md@@ -1,6 +1,6 @@---name: create-mod-description: Scaffold a Claude Code mod (a plugin with a function-hooks module) from ruflo's governed template β hybrid hooks.json with a classic fallback, a host adapter over literal $ calls, userConfig options, engine-kit tests and a tsconfig β then validate and test it. Use when the user asks for a mod, a function hook, or wants Claude Code to modify itself.+description: Scaffold a Claude Code mod (a plugin with a function-hooks module) from ruflo's governed template, then validate and test it. Use when the user asks for a mod, a function hook, or wants Claude Code to modify itself.argument-hint: "<mod-name>"allowed-tools: Bash Read Write Edit---diff --git a/plugins/ruflo-ruos/.claude-plugin/plugin.json b/plugins/ruflo-ruos/.claude-plugin/plugin.jsonindex cbeb83adc5..cd2d2c74be 100644--- a/plugins/ruflo-ruos/.claude-plugin/plugin.json+++ b/plugins/ruflo-ruos/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-ruos","description": "Run ruflo swarm agents on your own ruOS cloud desktops via the tenant-authenticated fleet MCP or per-tenant SSH; optional and removable (ADR-405)",- "version": "0.1.1",+ "version": "0.1.2","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-ruos/CHANGELOG.md b/plugins/ruflo-ruos/CHANGELOG.mdindex 97e20830c4..2dc70f4e4d 100644--- a/plugins/ruflo-ruos/CHANGELOG.md+++ b/plugins/ruflo-ruos/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.1.2 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.1.1 β 2026-10-05- feat: mod part β ruos status segment via $.ruflo.segment- feat: JobTransport (exec-poll + ruOS jobs API), live-run fixes, ruOS design rulesdiff --git a/plugins/ruflo-ruos/scripts/smoke.sh b/plugins/ruflo-ruos/scripts/smoke.shindex 182b4b41e0..8e845f7ce3 100755--- a/plugins/ruflo-ruos/scripts/smoke.sh+++ b/plugins/ruflo-ruos/scripts/smoke.sh@@ -9,9 +9,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json is valid and declares 0.1.1"+step "1. plugin.json is valid and declares 0.1.2"v=$(node -e 'const p=require(process.argv[1]);if(p.name!=="ruflo-ruos")process.exit(1);console.log(p.version)' "$ROOT/.claude-plugin/plugin.json" 2>/dev/null)-[[ "$v" == "0.1.1" ]] && ok || bad "got '$v'"+[[ "$v" == "0.1.2" ]] && ok || bad "got '$v'"step "2. commands, skill and agent present with frontmatter"miss=""diff --git a/plugins/ruflo-ruos/skills/ruos-host-run/SKILL.md b/plugins/ruflo-ruos/skills/ruos-host-run/SKILL.mdindex 5a071647b6..d56b13c639 100644--- a/plugins/ruflo-ruos/skills/ruos-host-run/SKILL.md+++ b/plugins/ruflo-ruos/skills/ruos-host-run/SKILL.md@@ -1,6 +1,6 @@---name: ruos-host-run-description: Run a ruflo swarm agent on the user's own ruOS cloud desktop through the tenant-authenticated fleet MCP, stream its output, and record it in ruflo's swarm state. Use when the user wants an agent to run on a ruOS desktop, asks to "run this on my ruOS desktop", or places a swarm agent on a remote ruOS host.+description: Run a ruflo swarm agent on the user's own ruOS cloud desktop through the fleet MCP, stream its output and record it in swarm state. Use when the user wants an agent to run on a ruOS desktop or places a swarm agent on a remote ruOS host.argument-hint: "<desktop> <task>"allowed-tools: Bash(node *) mcp__ruos__desktop_status mcp__ruos__desktop_exec mcp__ruos__desktop_keepawake mcp__ruos__desktop_start mcp__ruos__llm_route_get---diff --git a/plugins/ruflo-rure/.claude-plugin/plugin.json b/plugins/ruflo-rure/.claude-plugin/plugin.jsonindex cb3e4284b4..9c935804c2 100644--- a/plugins/ruflo-rure/.claude-plugin/plugin.json+++ b/plugins/ruflo-rure/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-rure","description": "Statically analyze a local JS/Electron app you're authorized to study, with file+line evidence, and build an original reimplementation of what you find. Experimental. Phase 1: JS/Electron static analysis. Phase 2: native ELF/PE/Mach-O container inspection. Phase 3: .NET assembly metadata. Phase 4: EVM bytecode inventory. Phase 5: CUDA fatbin, cubin and PTX inventory. Parses, never executes. Dynamic analysis is a future phase, not yet implemented.",- "version": "0.6.0",+ "version": "0.6.1","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-rure/CHANGELOG.md b/plugins/ruflo-rure/CHANGELOG.mdindex d55bedf3cf..9ea24cc592 100644--- a/plugins/ruflo-rure/CHANGELOG.md+++ b/plugins/ruflo-rure/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets. Built from git history; see `git log -- plugins/ruflo-rure` for the full record.+## 0.6.1 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.6.0 β 2026-10-24- feat: `rure_cuda_launch_plan` (ruflo ADR-501, ruflo CLI 3.62.0): the step from "what the CUDA program contains" to "what is worth measuring". For each kernel of each architecture it works out theoretical occupancy at a set of block sizes, which resource limits it, the best block sizes and a few flags, and with `includeHarness` returns the text of a CUDA driver-API program the person can build to check the numbers on their own GPU (and, with arguments they supply, to time launches). The occupancy model reproduces NVIDIA's `cuda_occupancy.h` on every combination checked for sm_70, 75, 80, 86, 87, 89, 90 and 120, and equals the CUDA driver on real devices (see the ADR for which). Read-only: nothing is compiled, loaded or run by ruRE, and nothing from the analyzed file reaches the harness text.diff --git a/plugins/ruflo-rure/skills/analyze-app/SKILL.md b/plugins/ruflo-rure/skills/analyze-app/SKILL.mdindex 59b63690be..a914413273 100644--- a/plugins/ruflo-rure/skills/analyze-app/SKILL.md+++ b/plugins/ruflo-rure/skills/analyze-app/SKILL.md@@ -1,6 +1,6 @@---name: analyze-app-description: Statically inspects a local JavaScript or Electron application you're authorized to study, explains how a feature works with file+line evidence, and helps build an equivalent for your own project. Use when the user wants to understand or reimplement a feature from an app whose source they have (their own app, a licensed target, or an explicit research/CTF context). Parses only β never executes the target.+description: Statically inspect a local JavaScript or Electron app you are authorized to study and explain how a feature works with file+line evidence. Use when the user wants to understand or reimplement a feature from an app whose source they have. Parses only; never executes the target.allowed-tools: Read---diff --git a/plugins/ruflo-ruvector/.claude-plugin/plugin.json b/plugins/ruflo-ruvector/.claude-plugin/plugin.jsonindex 6c53178a3f..db082e73b7 100644--- a/plugins/ruflo-ruvector/.claude-plugin/plugin.json+++ b/plugins/ruflo-ruvector/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-ruvector","description": "Self-learning vector database via npx [email protected] β HNSW, adaptive LoRA embeddings, code-graph clustering, hooks routing, brain/SONA, 103 MCP tools. As a mod (ADR-445): a write guard that keeps secrets out of the vector store and shared brain, /ruvector-mod, and a status file the console shows",- "version": "0.3.7",+ "version": "0.3.8","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-ruvector/CHANGELOG.md b/plugins/ruflo-ruvector/CHANGELOG.mdindex e8b254953b..0e032d02c9 100644--- a/plugins/ruflo-ruvector/CHANGELOG.md+++ b/plugins/ruflo-ruvector/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; older versions: `git log -- plugins/ruflo-ruvector`.+## 0.3.8 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.3.7 β 2026-10-14- fix: `vector-hyperbolic` no longer claims a normalised text embedding encodes depth (every unit-norm vector lands at the same radius); it now says to set radius from depth and measure on your own tree (#4159, thanks @ronmikailov)diff --git a/plugins/ruflo-ruvector/README.md b/plugins/ruflo-ruvector/README.mdindex fc32c4535e..830ed58da1 100644--- a/plugins/ruflo-ruvector/README.md+++ b/plugins/ruflo-ruvector/README.md@@ -54,6 +54,14 @@ Key tool categories: hooks routing, AST analysis, diff classification, coverage| `vector-cluster` | `/vector-cluster <files...>` | Spectral/Louvain code-graph clustering via `hooks graph-cluster` || `vector-hyperbolic` | `/vector-hyperbolic <text>` | Standard ONNX embed + Poincare projection in user code |+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. First-run setup installs packages, so it waits for you to ask.++| Skill | Run it with |+|---|---|+| vector-setup | `/ruflo-ruvector:vector-setup` |+## Commands (`/vector` slash command)The full surface is documented in `commands/vector.md` (80+ subcommands). Quick reference:diff --git a/plugins/ruflo-ruvector/commands/vector.md b/plugins/ruflo-ruvector/commands/vector.mdindex 8557aaafde..f27b725880 100644--- a/plugins/ruflo-ruvector/commands/vector.md+++ b/plugins/ruflo-ruvector/commands/vector.md@@ -142,7 +142,7 @@ Usage: /vector <subcommand> [options]## Setup helper-For first-run users hitting `ONNX WASM files not bundled`, `Brain commands require @ruvector/pi-brain`, or `SONA not available`, invoke the `vector-setup` skill: `/vector-setup`.+For first-run users hitting `ONNX WASM files not bundled`, `Brain commands require @ruvector/pi-brain`, or `SONA not available`, ask the user to run `/ruflo-ruvector:vector-setup` (it installs packages, so it is user-invoked: ADR-507).## MCP server (91 tools)diff --git a/plugins/ruflo-ruvector/skills/vector-setup/SKILL.md b/plugins/ruflo-ruvector/skills/vector-setup/SKILL.mdindex 3183ab5819..48be6d1dbe 100644--- a/plugins/ruflo-ruvector/skills/vector-setup/SKILL.md+++ b/plugins/ruflo-ruvector/skills/vector-setup/SKILL.md@@ -3,6 +3,7 @@ name: vector-setupdescription: First-run setup for [email protected] β installs ONNX/Brain/SONA add-ons, registers the MCP server, and verifies the install via `doctor`argument-hint: "[--full]"allowed-tools: Bash Read+disable-model-invocation: true---# Vector Setupdiff --git a/plugins/ruflo-testgen/.claude-plugin/plugin.json b/plugins/ruflo-testgen/.claude-plugin/plugin.jsonindex 9ab6a09b39..f2ffe29acc 100644--- a/plugins/ruflo-testgen/.claude-plugin/plugin.json+++ b/plugins/ruflo-testgen/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-testgen","description": "Test gap detection, coverage analysis, and automated test generation β drives the testgaps background worker via hooks_worker-dispatch; SPARC Refinement-phase canonical owner",- "version": "0.3.3",+ "version": "0.3.4","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-testgen/CHANGELOG.md b/plugins/ruflo-testgen/CHANGELOG.mdindex 12ae7dafbd..091b378d91 100644--- a/plugins/ruflo-testgen/CHANGELOG.md+++ b/plugins/ruflo-testgen/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; this is the whole history.+## 0.3.4 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.3.3 β 2026-10-12- fix: tdd-workflow reports the observed validation outcome to `hooks_post-task` instead of a literal `success: true` (#4085)diff --git a/plugins/ruflo-testgen/scripts/smoke.sh b/plugins/ruflo-testgen/scripts/smoke.shindex 14d5c8772f..4a46a55ade 100755--- a/plugins/ruflo-testgen/scripts/smoke.sh+++ b/plugins/ruflo-testgen/scripts/smoke.sh@@ -7,9 +7,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.3.3 with new keywords"+step "1. plugin.json declares 0.3.4 with new keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.3.3" ]]; then bad "expected 0.3.3, got '$v'"; else+if [[ "$v" != "0.3.4" ]]; then bad "expected 0.3.4, got '$v'"; elsemiss=""for k in mcp tdd-london-school coverage-routing; dogrep -q "\"$k\"" "$ROOT/.claude-plugin/plugin.json" || miss="$miss $k"diff --git a/plugins/ruflo-testgen/skills/tdd-repair/SKILL.md b/plugins/ruflo-testgen/skills/tdd-repair/SKILL.mdindex dab66c556a..b5d368487e 100644--- a/plugins/ruflo-testgen/skills/tdd-repair/SKILL.md+++ b/plugins/ruflo-testgen/skills/tdd-repair/SKILL.md@@ -1,6 +1,6 @@---name: tdd-repair-description: Test-Driven Repair β given a failing test, spawn a bounded headless `claude -p` (Read/Edit/Bash only) that makes the test pass without modifying it. Modeled on agent-harness-generator's ADR-175 Test-Driven Repair mode. Bounded cost via --max-budget-usd, bounded capability via --allowedTools. Closes the loop the TDD plugins didn't β we generate tests, this fixes the code to satisfy them.+description: "Test-Driven Repair: given a failing test, spawn a bounded headless `claude -p` (Read/Edit/Bash only, --max-budget-usd capped) that makes the test pass without modifying it. Use when a failing test should drive the fix to the code under test."argument-hint: "--repo <path> --test <path> --test-command <cmd> [--max-attempts 1] [--budget 5.00] [--model haiku] [--confirm]"allowed-tools: Bash---diff --git a/plugins/ruflo-workflows/.claude-plugin/plugin.json b/plugins/ruflo-workflows/.claude-plugin/plugin.jsonindex 4c357ad54c..79ed9d028b 100644--- a/plugins/ruflo-workflows/.claude-plugin/plugin.json+++ b/plugins/ruflo-workflows/.claude-plugin/plugin.json@@ -1,7 +1,7 @@{"name": "ruflo-workflows","description": "Persisted workflows, native orchestration and discovery of the optional managed cloud runtime.",- "version": "0.7.0",+ "version": "0.7.1","author": {"name": "ruvnet","url": "https://github.com/ruvnet"diff --git a/plugins/ruflo-workflows/CHANGELOG.md b/plugins/ruflo-workflows/CHANGELOG.mdindex 63732f4c8e..dfece36c50 100644--- a/plugins/ruflo-workflows/CHANGELOG.md+++ b/plugins/ruflo-workflows/CHANGELOG.md@@ -2,6 +2,9 @@Newest first. One `## <version> β <date>` heading per version, then `feat:`, `fix:`, `breaking:` and `chore:` bullets (ADR-478). Built from git history; older versions: `git log -- plugins/ruflo-workflows`.+## 0.7.1 β 2026-10-29+- chore: operational skills are user-invoked only (`disable-model-invocation: true`) and long skill/command descriptions are cut to one sentence, so they cost less prompt on every turn (ADR-507, #4588). Every skill is still available as `/<skill>`.+## 0.7.0 β 2026-10-24- feat: managed workflow capability discovery, bounded local mod policy and console policy metadata- fix: ship the managed workflow integration through a versioned plugin updatediff --git a/plugins/ruflo-workflows/README.md b/plugins/ruflo-workflows/README.mdindex 18a35888c9..02915bc0ed 100644--- a/plugins/ruflo-workflows/README.md+++ b/plugins/ruflo-workflows/README.md@@ -30,6 +30,15 @@ Neither subsumes the other β see [Choosing a surface](#choosing-a-surface).- `workflow-create` -- Author MCP workflow templates **or** native `.claude/workflows/*.js` orchestration scripts- `workflow-run` -- Execute/manage MCP workflows **or** invoke + resume native workflows+## User-invoked skills (ADR-507)++These skills carry `disable-model-invocation: true`: Claude does not pick them on its own, and they cost nothing in the per-turn skill listing. Type them to run them. The GAIA submission flow publishes to an external leaderboard; the comparison is a report.++| Skill | Run it with |+|---|---|+| gaia-architecture-comparison | `/ruflo-workflows:gaia-architecture-comparison` |+| gaia-submission | `/ruflo-workflows:gaia-submission` |+## Compatibility- **CLI:** pinned to `@claude-flow/cli` v3.6 major+minor.diff --git a/plugins/ruflo-workflows/scripts/smoke.sh b/plugins/ruflo-workflows/scripts/smoke.shindex 9d9e1f0d71..0a93c3700a 100755--- a/plugins/ruflo-workflows/scripts/smoke.sh+++ b/plugins/ruflo-workflows/scripts/smoke.sh@@ -7,9 +7,9 @@ step() { printf "β %s ... " "$1"; }ok() { printf "PASS\n"; PASS=$((PASS+1)); }bad() { printf "FAIL: %s\n" "$1"; FAIL=$((FAIL+1)); }-step "1. plugin.json declares 0.7.0 with lifecycle keywords"+step "1. plugin.json declares 0.7.1 with lifecycle keywords"v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)-if [[ "$v" != "0.7.0" ]]; then bad "expected 0.7.0, got '$v'"; else+if [[ "$v" != "0.7.1" ]]; then bad "expected 0.7.1, got '$v'"; elsemiss=""for k in mcp workflow-templates pause-resume lifecycle; dogrep -q "\"$k\"" "$ROOT/.claude-plugin/plugin.json" || miss="$miss $k"diff --git a/plugins/ruflo-workflows/skills/gaia-architecture-comparison/SKILL.md b/plugins/ruflo-workflows/skills/gaia-architecture-comparison/SKILL.mdindex 71e2b69082..df864e333e 100644--- a/plugins/ruflo-workflows/skills/gaia-architecture-comparison/SKILL.md+++ b/plugins/ruflo-workflows/skills/gaia-architecture-comparison/SKILL.md@@ -3,6 +3,7 @@ name: gaia-architecture-comparisondescription: Side-by-side comparison of ruflo vs HAL vs other GAIA harnesses β capability gaps, design decisions, and improvement roadmapargument-hint: "[--focus=tools|routing|memory|cost]"allowed-tools: Bash Read mcp__plugin_ruflo-core_ruflo__memory_search mcp__plugin_ruflo-core_ruflo__memory_store+disable-model-invocation: true---# GAIA Architecture Comparison Skilldiff --git a/plugins/ruflo-workflows/skills/gaia-submission/SKILL.md b/plugins/ruflo-workflows/skills/gaia-submission/SKILL.mdindex 786c375efe..d46b13083a 100644--- a/plugins/ruflo-workflows/skills/gaia-submission/SKILL.md+++ b/plugins/ruflo-workflows/skills/gaia-submission/SKILL.md@@ -3,6 +3,7 @@ name: gaia-submissiondescription: Walk through a complete GAIA benchmarkβsubmit flow β from key resolution through HAL-compatible package generationargument-hint: "[level] [limit] [models]"allowed-tools: Bash mcp__plugin_ruflo-core_ruflo__memory_store mcp__plugin_ruflo-core_ruflo__memory_search mcp__plugin_ruflo-core_ruflo__memory_list mcp__plugin_ruflo-core_ruflo__hooks_post-task mcp__plugin_ruflo-core_ruflo__hooks_pre-task+disable-model-invocation: true---# GAIA Submission Skilldiff --git a/scripts/audit-skill-frontmatter.mjs b/scripts/audit-skill-frontmatter.mjsindex 7f628e005a..10f7ec7434 100755--- a/scripts/audit-skill-frontmatter.mjs+++ b/scripts/audit-skill-frontmatter.mjs@@ -13,6 +13,7 @@// node scripts/audit-skill-frontmatter.mjs # all skills// node scripts/audit-skill-frontmatter.mjs --format json # machine-readable// node scripts/audit-skill-frontmatter.mjs --only ruflo-cost-tracker+// node scripts/audit-skill-frontmatter.mjs --plugins-dir <dir> # audit another tree (tests)//// CHECKS per SKILL.md// 1. File has a `---` frontmatter block at the top.@@ -21,31 +22,67 @@// 4. `allowed-tools:` field present (security β no implicit "all tools").// 5. `allowed-tools:` is NOT a wildcard (`*`).// 6. `name:` value matches the directory name (drift guard).+// 7. `description:` is at most 300 characters (ADR-507 Β§2.3). Also checked+// for every plugins/*/commands/*.md, since commands share the per-turn+// listing. Multi-line (`>`, `|`) descriptions are measured unfolded.+// 8. A skill with `disable-model-invocation: true` is named in a+// "User-invoked" section of its plugin README (ADR-507 Β§3), so a person+// can find the `/<skill>` the model no longer discovers on its own.//// EXIT CODES// 0 no violations// 1 at least one violation found// 2 scan error (no plugins dir)-import { readdirSync, statSync, readFileSync } from 'node:fs';-import { dirname, join, basename } from 'node:path';+import { readdirSync, statSync, readFileSync, realpathSync } from 'node:fs';+import { dirname, join, resolve } from 'node:path';import { fileURLToPath } from 'node:url';+import { parseFrontmatter as parseFullFrontmatter, isUserInvokedOnly, DESCRIPTION_CEILING } from './measure-plugin-footprint.mjs';++// ADR-507 Β§3: the text of the README section(s) whose heading says "user-invoked".+export function userInvokedSection(readme) {+ const lines = String(readme || '').split(/\r?\n/);+ const out = [];+ let level = 0;+ for (const line of lines) {+ const h = /^(#{1,6})\s+(.*)$/.exec(line);+ if (h) {+ if (/user[- ]invoked/i.test(h[2])) { level = h[1].length; continue; }+ if (level && h[1].length <= level) level = 0;+ }+ if (level) out.push(line);+ }+ return out.join('\n');+}++export function isListedAsUserInvoked(readme, name) {+ const esc = name.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');+ return new RegExp(`(^|[^\\w-])${esc}($|[^\\w-])`).test(userInvokedSection(readme));+}++export function descriptionTooLong(description) {+ return typeof description === 'string' && description.length > DESCRIPTION_CEILING;+}const SCRIPTS_DIR = dirname(fileURLToPath(import.meta.url));const REPO_ROOT = dirname(SCRIPTS_DIR);-const PLUGINS_DIR = join(REPO_ROOT, 'plugins');const ARGS = (() => {- const a = { format: 'table', only: null };+ const a = { format: 'table', only: null, pluginsDir: join(REPO_ROOT, 'plugins') };for (let i = 2; i < process.argv.length; i++) {const v = process.argv[i];if (v === '--format') a.format = process.argv[++i];else if (v === '--only') {a.only = new Set((process.argv[++i] || '').split(',').map((s) => s.trim()).filter(Boolean));+ } else if (v === '--plugins-dir') {+ const d = process.argv[++i];+ if (!d) { console.error('audit-skill-frontmatter: --plugins-dir needs a directory'); process.exit(2); }+ a.pluginsDir = resolve(d);}}return a;})();+const PLUGINS_DIR = ARGS.pluginsDir;function discoverSkills() {const out = [];@@ -118,9 +155,57 @@ function auditSkill(skill) {detail: `frontmatter name="${fm.name}" but directory is "${skill.skillDir}"`,});}+ // ADR-507: measure the unfolded description and the user-invoked flag with+ // the full parser β the flat one above reads a `>` block as the literal `>`.+ const full = parseFullFrontmatter(src) || {};+ if (descriptionTooLong(full.description)) {+ violations.push({+ check: 'description length',+ detail: `${full.description.length} chars > ${DESCRIPTION_CEILING} (ADR-507: one sentence of what it does and when to use it)`,+ });+ }+ if (isUserInvokedOnly(full)) {+ const name = fm.name || skill.skillDir;+ let readme = '';+ try { readme = readFileSync(join(PLUGINS_DIR, skill.plugin, 'README.md'), 'utf-8'); } catch {}+ if (!isListedAsUserInvoked(readme, name)) {+ violations.push({+ check: 'user-invoked skill listed in README',+ detail: `disable-model-invocation: true but "${name}" is not named under a "User-invoked" heading in plugins/${skill.plugin}/README.md`,+ });+ }+ }return violations;}+// ADR-507 Β§2.3: commands share the per-turn listing, so their descriptions get the same ceiling.+function auditCommands() {+ const out = [];+ let plugins;+ try { plugins = readdirSync(PLUGINS_DIR); } catch { return out; }+ for (const plugin of plugins) {+ if (ARGS.only && !ARGS.only.has(plugin)) continue;+ const dir = join(PLUGINS_DIR, plugin, 'commands');+ let entries;+ try { entries = readdirSync(dir); } catch { continue; }+ for (const e of entries) {+ if (!e.endsWith('.md')) continue;+ let src;+ try { src = readFileSync(join(dir, e), 'utf-8'); } catch { continue; }+ const fm = parseFullFrontmatter(src);+ if (fm && descriptionTooLong(fm.description)) {+ out.push({+ plugin,+ skill: `commands/${e}`,+ file: join(dir, e).replace(REPO_ROOT + '/', ''),+ violations: [{ check: 'description length', detail: `${fm.description.length} chars > ${DESCRIPTION_CEILING} (ADR-507)` }],+ });+ }+ }+ }+ return out;+}+function main() {const skills = discoverSkills();if (skills.length === 0) {@@ -139,6 +224,7 @@ function main() {});}}+ findings.push(...auditCommands());if (ARGS.format === 'json') {console.log(JSON.stringify({@@ -159,6 +245,8 @@ function main() {console.log(' - name / description / allowed-tools all present and non-empty');console.log(' - no wildcard allowed-tools');console.log(' - name matches enclosing directory');+ console.log(` - skill and command descriptions within ${DESCRIPTION_CEILING} chars`);+ console.log(' - every user-invoked (disable-model-invocation) skill named in its plugin README');} else {console.log(`β Found ${findings.length} skill(s) with frontmatter violations:`);console.log('');@@ -175,4 +263,12 @@ function main() {process.exit(findings.length > 0 ? 1 : 0);}-main();+// Run main() only when executed, not when imported. Node realpaths+// import.meta.url for the entry module but leaves argv[1] as typed, so compare+// realpaths: a symlinked invocation path must still run the gate.+export function isEntryPoint(metaUrl, argv1 = process.argv[1]) {+ if (!argv1) return false;+ try { return realpathSync(argv1) === realpathSync(fileURLToPath(metaUrl)); } catch { return false; }+}++if (isEntryPoint(import.meta.url)) main();diff --git a/scripts/measure-plugin-footprint.mjs b/scripts/measure-plugin-footprint.mjsnew file mode 100644index 0000000000..3771e585b0--- /dev/null+++ b/scripts/measure-plugin-footprint.mjs@@ -0,0 +1,262 @@+#!/usr/bin/env node+// measure-plugin-footprint β ADR-507 Β§2.1. How much of every turn's context the+// ruflo plugin suite costs before the model does anything.+//+// Claude Code keeps the name and description of every skill, command and agent+// the model can invoke in context on every turn. This script reports, per+// plugin under plugins/:+// - skills, commands and agents (and how many are user-invoked only, i.e.+// carry `disable-model-invocation: true` and so leave the model's listing);+// - the characters each contributes to the per-turn listing (name plus+// description, model-visible entries only);+// - an estimated token count (characters / 4).+// It also lists every skill or command description longer than the ADR-507+// ceiling (300 characters).+//+// USAGE+// node scripts/measure-plugin-footprint.mjs # table to stdout+// node scripts/measure-plugin-footprint.mjs --format json # machine-readable+// node scripts/measure-plugin-footprint.mjs --write [--label before|after]+// writes v3/docs/validation/plugin-footprint-<date>[-<label>].md+// node scripts/measure-plugin-footprint.mjs --ref <git-ref> # measure a commit instead of the tree+//+// The token figure is an estimate, not a tokenizer count; it is meant for+// before/after comparison on the same estimator.+//+// EXIT CODES+// 0 measured+// 2 scan error (no plugins dir, bad args)++import { readdirSync, readFileSync, statSync, writeFileSync, mkdirSync, realpathSync } from 'node:fs';+import { execFileSync } from 'node:child_process';+import { dirname, join } from 'node:path';+import { fileURLToPath } from 'node:url';++const REPO_ROOT = dirname(dirname(fileURLToPath(import.meta.url)));+export const DESCRIPTION_CEILING = 300;++function parseArgs(argv) {+ const a = { format: 'table', write: false, label: '', ref: null };+ for (let i = 2; i < argv.length; i++) {+ const v = argv[i];+ if (v === '--format') a.format = argv[++i];+ else if (v === '--write') a.write = true;+ else if (v === '--label') a.label = String(argv[++i] || '').replace(/[^\w-]/g, '');+ else if (v === '--ref') a.ref = argv[++i];+ else if (v === '--help' || v === '-h') { a.help = true; }+ else { console.error(`measure-plugin-footprint: unknown argument ${v}`); process.exit(2); }+ }+ if (a.ref !== null && !/^[\w./~^-]+$/.test(a.ref)) { console.error('measure-plugin-footprint: bad --ref'); process.exit(2); }+ return a;+}++// ββ frontmatter ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ+// Flat YAML subset: `key: value`, quoted values, block scalars (`>`, `|`, with+// chomping indicators) and plain multi-line continuations. Enough for SKILL.md,+// command and agent frontmatter; nested maps are skipped.+export function parseFrontmatter(src) {+ const lines = src.split(/\r?\n/);+ if ((lines[0] || '').trim() !== '---') return null;+ let end = -1;+ for (let i = 1; i < lines.length; i++) if (lines[i].trim() === '---') { end = i; break; }+ if (end < 0) return null;+ const body = lines.slice(1, end);+ const out = {};+ for (let i = 0; i < body.length; i++) {+ const m = /^([\w-]+):[ \t]*(.*)$/.exec(body[i]);+ if (!m) continue;+ const key = m[1];+ let raw = m[2];+ const cont = [];+ let j = i + 1;+ while (j < body.length && (/^[ \t]+\S/.test(body[j]) || body[j].trim() === '')) { cont.push(body[j]); j++; }+ while (cont.length && cont[cont.length - 1].trim() === '') cont.pop();+ if (/^[>|][+-]?$/.test(raw.trim())) {+ const folded = raw.trim()[0] === '>';+ const parts = cont.map((l) => l.trim());+ raw = folded ? parts.join(' ').replace(/\s+/g, ' ') : parts.join('\n');+ } else if (cont.length) {+ raw = [raw, ...cont.map((l) => l.trim())].join(' ');+ }+ raw = raw.trim();+ if ((raw.startsWith('"') && raw.endsWith('"')) || (raw.startsWith("'") && raw.endsWith("'"))) {+ const q = raw[0];+ raw = raw.slice(1, -1);+ raw = q === '"' ? raw.replace(/\\"/g, '"').replace(/\\\\/g, '\\') : raw.replace(/''/g, "'");+ }+ out[key] = raw;+ i = j - 1;+ }+ return out;+}++export function isUserInvokedOnly(fm) {+ return !!fm && String(fm['disable-model-invocation'] || '').trim().toLowerCase() === 'true';+}++// ββ sources: the working tree or a git ref ββββββββββββββββββββββββββββββββββ+function treeSource() {+ const pluginsDir = join(REPO_ROOT, 'plugins');+ const files = [];+ const walk = (rel) => {+ let entries;+ try { entries = readdirSync(join(REPO_ROOT, rel)); } catch { return; }+ for (const e of entries) {+ const r = `${rel}/${e}`;+ let s;+ try { s = statSync(join(REPO_ROOT, r)); } catch { continue; }+ if (s.isDirectory()) { if (e !== 'node_modules' && !e.startsWith('.')) walk(r); }+ else files.push(r);+ }+ };+ try { statSync(pluginsDir); } catch { return null; }+ for (const p of readdirSync(pluginsDir)) {+ for (const sub of ['skills', 'commands', 'agents']) walk(`plugins/${p}/${sub}`);+ }+ return { files, read: (rel) => readFileSync(join(REPO_ROOT, rel), 'utf-8') };+}++function refSource(ref) {+ const git = (args) => execFileSync('git', args, { cwd: REPO_ROOT, encoding: 'utf-8', maxBuffer: 64 << 20 });+ const files = git(['ls-tree', '-r', '--name-only', ref, '--', 'plugins']).split('\n').filter(Boolean);+ return { files, read: (rel) => git(['show', `${ref}:${rel}`]) };+}++// skills/<name>/SKILL.md, commands/**/*.md, agents/**/*.md β what Claude Code loads.+function classify(rel) {+ const m = /^plugins\/([^/]+)\/(skills|commands|agents)\/(.+)$/.exec(rel);+ if (!m) return null;+ const [, plugin, kind, rest] = m;+ if (kind === 'skills') {+ const s = /^([^/]+)\/SKILL\.md$/.exec(rest);+ return s ? { plugin, kind: 'skill', fallbackName: s[1] } : null;+ }+ if (!rest.endsWith('.md') || /(^|\/)README\.md$/i.test(rest)) return null;+ const base = rest.replace(/\.md$/, '').split('/').pop();+ return { plugin, kind: kind === 'commands' ? 'command' : 'agent', fallbackName: base };+}++export function measure(source) {+ const plugins = new Map();+ const over = [];+ const userInvoked = [];+ const get = (p) => {+ if (!plugins.has(p)) {+ plugins.set(p, {+ plugin: p, skills: 0, commands: 0, agents: 0, userInvokedSkills: 0, userInvokedCommands: 0,+ listingChars: 0, skillChars: 0, commandChars: 0, agentChars: 0,+ });+ }+ return plugins.get(p);+ };+ for (const rel of source.files) {+ const c = classify(rel);+ if (!c) continue;+ let fm;+ try { fm = parseFrontmatter(source.read(rel)); } catch { fm = null; }+ const name = (fm && fm.name) || c.fallbackName;+ const description = (fm && fm.description) || '';+ const row = get(c.plugin);+ const hidden = c.kind !== 'agent' && isUserInvokedOnly(fm);+ if (c.kind === 'skill') { row.skills++; if (hidden) row.userInvokedSkills++; }+ else if (c.kind === 'command') { row.commands++; if (hidden) row.userInvokedCommands++; }+ else row.agents++;+ if (hidden) userInvoked.push({ plugin: c.plugin, kind: c.kind, name, file: rel });+ if (c.kind !== 'agent' && description.length > DESCRIPTION_CEILING) {+ over.push({ plugin: c.plugin, kind: c.kind, name, chars: description.length, file: rel });+ }+ if (hidden) continue;+ const chars = name.length + description.length;+ row.listingChars += chars;+ row[`${c.kind}Chars`] += chars;+ }+ const rows = [...plugins.values()].sort((a, b) => b.listingChars - a.listingChars || a.plugin.localeCompare(b.plugin));+ const totals = rows.reduce((t, r) => {+ for (const k of Object.keys(t)) t[k] += r[k];+ return t;+ }, { skills: 0, commands: 0, agents: 0, userInvokedSkills: 0, userInvokedCommands: 0, listingChars: 0, skillChars: 0, commandChars: 0, agentChars: 0 });+ const tokens = (n) => Math.round(n / 4);+ return {+ rows: rows.map((r) => ({ ...r, estTokens: tokens(r.listingChars) })),+ totals: { ...totals, estTokens: tokens(totals.listingChars) },+ overCeiling: over.sort((a, b) => b.chars - a.chars),+ userInvoked: userInvoked.sort((a, b) => (a.plugin + a.name).localeCompare(b.plugin + b.name)),+ };+}++function head(ref) {+ try { return execFileSync('git', ['rev-parse', '--short', ref || 'HEAD'], { cwd: REPO_ROOT, encoding: 'utf-8' }).trim(); } catch { return 'unknown'; }+}++export function renderMarkdown(r, { date, label, commit, dirty }) {+ const t = r.totals;+ const L = [];+ L.push(`# Plugin footprint${label ? ` (${label})` : ''}, ${date}`, '');+ L.push(`Measured by \`node scripts/measure-plugin-footprint.mjs\` (ADR-507 Β§2.1) at \`${commit}\`${dirty ? ' plus uncommitted changes' : ''}.`, '');+ L.push('The listing is what Claude Code keeps in context on every turn: the name and description of each skill, command and agent the model can invoke. A skill or command with `disable-model-invocation: true` stays runnable as `/<name>` but leaves that listing, so it is not counted. Tokens are estimated as characters Γ· 4.', '');+ L.push('## Totals', '');+ L.push('| | Count |', '|---|---:|');+ L.push(`| Skills | ${t.skills} (${t.userInvokedSkills} user-invoked only) |`);+ L.push(`| Commands | ${t.commands} (${t.userInvokedCommands} user-invoked only) |`);+ L.push(`| Agents | ${t.agents} |`);+ L.push(`| Listing characters | ${t.listingChars.toLocaleString('en-US')} (skills ${t.skillChars.toLocaleString('en-US')}, commands ${t.commandChars.toLocaleString('en-US')}, agents ${t.agentChars.toLocaleString('en-US')}) |`);+ L.push(`| Estimated tokens per turn | ${t.estTokens.toLocaleString('en-US')} |`);+ L.push(`| Skill/command descriptions over ${DESCRIPTION_CEILING} characters | ${r.overCeiling.length} |`, '');+ L.push('These totals assume every ruflo plugin is enabled at once; a session pays only for the plugins it enables.', '');+ L.push('## Per plugin', '');+ L.push('| Plugin | Skills (user-invoked) | Commands (user-invoked) | Agents | Listing chars | Est. tokens |', '|---|---:|---:|---:|---:|---:|');+ for (const row of r.rows) {+ L.push(`| ${row.plugin} | ${row.skills} (${row.userInvokedSkills}) | ${row.commands} (${row.userInvokedCommands}) | ${row.agents} | ${row.listingChars.toLocaleString('en-US')} | ${row.estTokens.toLocaleString('en-US')} |`);+ }+ L.push('');+ L.push(`## Descriptions over ${DESCRIPTION_CEILING} characters`, '');+ if (!r.overCeiling.length) L.push('None.', '');+ else {+ L.push('| Plugin | Kind | Name | Characters |', '|---|---|---|---:|');+ for (const o of r.overCeiling) L.push(`| ${o.plugin} | ${o.kind} | ${o.name} | ${o.chars} |`);+ L.push('');+ }+ L.push('## User-invoked only (`disable-model-invocation: true`)', '');+ if (!r.userInvoked.length) L.push('None.', '');+ else {+ L.push('| Plugin | Kind | Name |', '|---|---|---|');+ for (const u of r.userInvoked) L.push(`| ${u.plugin} | ${u.kind} | ${u.name} |`);+ L.push('');+ }+ return L.join('\n');+}++function main() {+ const args = parseArgs(process.argv);+ if (args.help) { console.log(readFileSync(fileURLToPath(import.meta.url), 'utf-8').split('\n').slice(1, 30).join('\n')); return; }+ const source = args.ref ? refSource(args.ref) : treeSource();+ if (!source || !source.files.length) { console.error('measure-plugin-footprint: no plugins found'); process.exit(2); }+ const r = measure(source);+ if (args.format === 'json') { console.log(JSON.stringify(r, null, 2)); }+ else {+ const t = r.totals;+ console.log(`skills ${t.skills} (${t.userInvokedSkills} user-invoked) Β· commands ${t.commands} (${t.userInvokedCommands} user-invoked) Β· agents ${t.agents}`);+ console.log(`listing ${t.listingChars} chars β ${t.estTokens} tokens/turn Β· ${r.overCeiling.length} description(s) over ${DESCRIPTION_CEILING} chars`);+ for (const row of r.rows.slice(0, 15)) {+ console.log(` ${row.plugin.padEnd(28)} ${String(row.listingChars).padStart(6)} chars β${String(row.estTokens).padStart(5)} tok skills ${row.skills}/${row.userInvokedSkills}u cmds ${row.commands}/${row.userInvokedCommands}u agents ${row.agents}`);+ }+ }+ if (args.write) {+ const date = new Date().toISOString().slice(0, 10);+ const dirty = !args.ref && (() => { try { return execFileSync('git', ['status', '--porcelain', '--untracked-files=no', '--', 'plugins'], { cwd: REPO_ROOT, encoding: 'utf-8' }).trim() !== ''; } catch { return false; } })();+ const md = renderMarkdown(r, { date, label: args.label, commit: head(args.ref), dirty });+ const dir = join(REPO_ROOT, 'v3/docs/validation');+ mkdirSync(dir, { recursive: true });+ const file = join(dir, `plugin-footprint-${date}${args.label ? `-${args.label}` : ''}.md`);+ writeFileSync(file, md + '\n');+ console.error(`wrote ${file.replace(REPO_ROOT + '/', '')}`);+ }+}++// Compare realpaths: Node realpaths import.meta.url for the entry module but+// not argv[1], so a symlinked invocation path would otherwise skip main().+const isEntry = (() => {+ if (!process.argv[1]) return false;+ try { return realpathSync(process.argv[1]) === realpathSync(fileURLToPath(import.meta.url)); } catch { return false; }+})();+if (isEntry) main();diff --git a/v3/@claude-flow/cli/__tests__/plugin-footprint-adr507.test.ts b/v3/@claude-flow/cli/__tests__/plugin-footprint-adr507.test.tsnew file mode 100644index 0000000000..d9dbc7b7b1--- /dev/null+++ b/v3/@claude-flow/cli/__tests__/plugin-footprint-adr507.test.ts@@ -0,0 +1,166 @@+/**+ * ADR-507 β plugin footprint: the measure script, the 300-character+ * description ceiling, and the "user-invoked skills are listed in the README"+ * rule enforced by scripts/audit-skill-frontmatter.mjs.+ */+import { describe, expect, it } from 'vitest';+import { spawnSync } from 'node:child_process';+import { mkdirSync, mkdtempSync, rmSync, symlinkSync, writeFileSync } from 'node:fs';+import { tmpdir } from 'node:os';+import { dirname, join, resolve } from 'node:path';+import { fileURLToPath, pathToFileURL } from 'node:url';++const REPO = resolve(dirname(fileURLToPath(import.meta.url)), '../../../..');+const load = (rel: string) => import(pathToFileURL(resolve(REPO, rel)).href);++function source(files: Record<string, string>) {+ return { files: Object.keys(files), read: (rel: string) => files[rel] };+}++const skill = (name: string, description: string, extra = '') =>+ `---\nname: ${name}\ndescription: ${description}\nallowed-tools: Bash\n${extra}---\n\n# ${name}\n`;++describe('ADR-507 measure-plugin-footprint', () => {+ it('parses folded, quoted and plain multi-line descriptions unfolded', async () => {+ const { parseFrontmatter } = await load('scripts/measure-plugin-footprint.mjs');+ expect(parseFrontmatter('---\nname: a\ndescription: >\n one two\n three\nallowed-tools: Bash\n---\n').description).toBe('one two three');+ expect(parseFrontmatter('---\ndescription: "say: hi"\n---\n').description).toBe('say: hi');+ expect(parseFrontmatter("---\ndescription: it''s\n---\n").description).toBe("it''s");+ expect(parseFrontmatter("---\ndescription: 'it''s'\n---\n").description).toBe("it's");+ expect(parseFrontmatter('---\ndescription: first\n second\n---\n').description).toBe('first second');+ expect(parseFrontmatter('no frontmatter')).toBeNull();+ });++ it('drops a user-invoked skill from the listing but still counts it', async () => {+ const { measure } = await load('scripts/measure-plugin-footprint.mjs');+ const desc = 'Make a report. Use when asked for one.';+ const visible = measure(source({ 'plugins/p/skills/rep/SKILL.md': skill('rep', desc) }));+ const hidden = measure(source({ 'plugins/p/skills/rep/SKILL.md': skill('rep', desc, 'disable-model-invocation: true\n') }));+ expect(visible.totals.listingChars).toBe('rep'.length + desc.length);+ expect(hidden.totals.listingChars).toBe(0); // negative control: the flag is what removes it+ expect(hidden.totals.skills).toBe(1);+ expect(hidden.totals.userInvokedSkills).toBe(1);+ expect(hidden.userInvoked.map((u: { name: string }) => u.name)).toEqual(['rep']);+ expect(visible.totals.estTokens).toBe(Math.round(visible.totals.listingChars / 4));+ });++ it('flags descriptions over 300 characters and not at 300', async () => {+ const { measure, DESCRIPTION_CEILING } = await load('scripts/measure-plugin-footprint.mjs');+ expect(DESCRIPTION_CEILING).toBe(300);+ const r = measure(source({+ 'plugins/p/skills/at/SKILL.md': skill('at', 'x'.repeat(300)),+ 'plugins/p/skills/over/SKILL.md': skill('over', 'x'.repeat(301)),+ 'plugins/p/commands/c.md': `---\nname: c\ndescription: ${'y'.repeat(301)}\n---\n`,+ }));+ expect(r.overCeiling.map((o: { name: string }) => o.name).sort()).toEqual(['c', 'over']);+ });++ it('counts agents and commands, and ignores files Claude Code does not load as skills', async () => {+ const { measure } = await load('scripts/measure-plugin-footprint.mjs');+ const r = measure(source({+ 'plugins/p/agents/helper.md': '---\nname: helper\ndescription: helps\n---\n',+ 'plugins/p/commands/go.md': '---\nname: go\ndescription: goes\n---\n',+ 'plugins/p/commands/README.md': '# readme',+ 'plugins/p/directory/skills/copy/SKILL.md': skill('copy', 'a submission copy'),+ 'plugins/p/skills/real/notes.md': 'not a skill',+ }));+ expect(r.totals).toMatchObject({ skills: 0, commands: 1, agents: 1 });+ expect(r.totals.listingChars).toBe('helper'.length + 'helps'.length + 'go'.length + 'goes'.length);+ });+});++describe('ADR-507 audit-skill-frontmatter rules', () => {+ it('accepts a user-invoked skill only when a "User-invoked" section names it', async () => {+ const { isListedAsUserInvoked } = await load('scripts/audit-skill-frontmatter.mjs');+ const readme = '# P\n\n## Skills\n\n- cost-report\n\n## User-invoked skills (ADR-507)\n\n| cost-anomaly | `/p:cost-anomaly` |\n\n## Commands\n\n- cost-burn\n';+ expect(isListedAsUserInvoked(readme, 'cost-anomaly')).toBe(true);+ expect(isListedAsUserInvoked(readme, 'cost-report')).toBe(false); // named, but not in the section+ expect(isListedAsUserInvoked(readme, 'cost-burn')).toBe(false); // after the section ended+ expect(isListedAsUserInvoked(readme, 'cost')).toBe(false); // no partial-name match+ expect(isListedAsUserInvoked('', 'cost-anomaly')).toBe(false);+ });++ it('applies the 300-character ceiling', async () => {+ const { descriptionTooLong } = await load('scripts/audit-skill-frontmatter.mjs');+ expect(descriptionTooLong('x'.repeat(300))).toBe(false);+ expect(descriptionTooLong('x'.repeat(301))).toBe(true);+ expect(descriptionTooLong(undefined)).toBe(false);+ });++ // Run the real CLI against a fixture tree, so the gate's wiring (not just+ // the helpers) is under test: disabling any check in auditSkill/auditCommands+ // fails one of these.+ function fixture(readmeRow: boolean) {+ const dir = mkdtempSync(join(tmpdir(), 'adr507-audit-'));+ const put = (rel: string, body: string) => {+ mkdirSync(dirname(join(dir, rel)), { recursive: true });+ writeFileSync(join(dir, rel), body);+ };+ const longFolded = ('word '.repeat(61)).trim(); // 304 chars once unfolded+ put('p/skills/long/SKILL.md', `---\nname: long\ndescription: >\n ${longFolded.slice(0, 149)}\n ${longFolded.slice(150)}\nallowed-tools: Bash\n---\n`);+ put('p/skills/hidden/SKILL.md', skill('hidden', 'A report. Use when asked.', 'disable-model-invocation: true\n'));+ put('p/skills/ok/SKILL.md', skill('ok', 'Fine. Use when fine.'));+ put('p/commands/big.md', `---\nname: big\ndescription: |\n ${'z'.repeat(200)}\n ${'z'.repeat(110)}\n---\n`);+ put('p/README.md', readmeRow ? '# P\n\n## User-invoked skills\n\n- `hidden`\n' : '# P\n\n## Skills\n\n- hidden\n');+ return dir;+ }+ const audit = (dir: string, script = 'scripts/audit-skill-frontmatter.mjs') =>+ spawnSync(process.execPath, [script, '--format', 'json', '--plugins-dir', dir], { cwd: REPO, encoding: 'utf8' });++ it('the CLI fails a fixture with a long folded description, an unlisted user-invoked skill and a long command', () => {+ const dir = fixture(false);+ try {+ const out = audit(dir);+ expect(out.status).toBe(1);+ const report = JSON.parse(out.stdout);+ expect(report.skillsScanned).toBe(3);+ const checks = (name: string) =>+ report.findings.find((f: { skill: string }) => f.skill === name)?.violations.map((v: { check: string }) => v.check) ?? [];+ expect(checks('long')).toContain('description length');+ expect(checks('hidden')).toContain('user-invoked skill listed in README');+ expect(checks('commands/big.md')).toContain('description length');+ expect(checks('ok')).toEqual([]);+ } finally { rmSync(dir, { recursive: true, force: true }); }+ });++ it('control: the README row clears the user-invoked finding, the length findings stay', () => {+ const dir = fixture(true);+ try {+ const report = JSON.parse(audit(dir).stdout);+ const flat = report.findings.flatMap((f: { skill: string; violations: { check: string }[] }) => f.violations.map((v) => `${f.skill}:${v.check}`));+ expect(flat.sort()).toEqual(['commands/big.md:description length', 'long:description length']);+ } finally { rmSync(dir, { recursive: true, force: true }); }+ });++ it('still runs the gate when invoked through a symlinked path', () => {+ const dir = fixture(false);+ const link = join(tmpdir(), `adr507-link-${process.pid}-${Date.now()}`);+ try {+ symlinkSync(REPO, link);+ for (const script of ['scripts/audit-skill-frontmatter.mjs', 'scripts/measure-plugin-footprint.mjs']) {+ const args = script.includes('audit') ? ['--format', 'json', '--plugins-dir', dir] : ['--format', 'json'];+ const out = spawnSync(process.execPath, [join(link, script), ...args], { encoding: 'utf8' });+ expect(out.stdout.trim().length, script).toBeGreaterThan(0); // main() ran+ if (script.includes('audit')) expect(out.status).toBe(1);+ else expect(JSON.parse(out.stdout).totals.skills).toBeGreaterThan(0);+ }+ } finally {+ rmSync(link, { force: true });+ rmSync(dir, { recursive: true, force: true });+ }+ });++ it('passes on the repository: no long descriptions, every user-invoked skill documented', async () => {+ const out = spawnSync(process.execPath, ['scripts/audit-skill-frontmatter.mjs', '--format', 'json'], { cwd: REPO, encoding: 'utf8' });+ const report = JSON.parse(out.stdout);+ expect(report.skillsScanned).toBeGreaterThan(100); // zero checks must not read as zero violations+ expect(report.findings).toEqual([]);+ expect(out.status).toBe(0);+ const { measure } = await load('scripts/measure-plugin-footprint.mjs');+ const tree = spawnSync('git', ['ls-files', 'plugins'], { cwd: REPO, encoding: 'utf8' }).stdout.split('\n').filter(Boolean);+ const { readFileSync } = await import('node:fs');+ const r = measure({ files: tree, read: (rel: string) => readFileSync(resolve(REPO, rel), 'utf8') });+ expect(r.overCeiling).toEqual([]);+ expect(r.totals.userInvokedSkills).toBeGreaterThan(0);+ });+});diff --git a/v3/docs/adr/ADR-507-plugin-footprint-skills-commands.md b/v3/docs/adr/ADR-507-plugin-footprint-skills-commands.mdindex 5efd41f668..190a7e1b21 100644--- a/v3/docs/adr/ADR-507-plugin-footprint-skills-commands.md+++ b/v3/docs/adr/ADR-507-plugin-footprint-skills-commands.md@@ -1,6 +1,6 @@# ADR 507: cutting the plugin suite's per-turn footprint-Status: Proposed (2026-10-29)+Status: Accepted (implemented in PR #4588; Β§2.4 closed, see Implementation notes)Issue: #4581 follow-up 4. Builds on: ADR-451 (`agentTrim`).@@ -58,3 +58,57 @@ If the engine exposes no such hook, `skillTrim` is not built and this section is- Footprint report before and after, committed.- Each changed SKILL.md still passes the plugin smoke contracts and the ADR-112 audit.- A smoke check confirms that every `disable-model-invocation: true` skill is still listed in its plugin README.++## 5. Implementation notes++**Measurement (Β§2.1).** `scripts/measure-plugin-footprint.mjs` reads `plugins/*/skills/<name>/SKILL.md`, `commands/**/*.md` and `agents/**/*.md` from the tree or from a git ref (`--ref`), and writes `v3/docs/validation/plugin-footprint-<date>[-<label>].md`. With every ruflo plugin enabled:++| | Before (origin/main) | After |+|---|---:|---:|+| Listing characters | 47,743 | 39,402 |+| Estimated tokens per turn | 11,936 | 9,851 |+| User-invoked skills | 0 | 27 |+| Descriptions over 300 characters | 20 | 0 |++That is about 2,085 tokens per turn with everything enabled. The largest drops: ruflo-cost-tracker 1,271 β 542 tokens, ruflo-metaharness 1,255 β 545. Reports: `v3/docs/validation/plugin-footprint-2026-10-29-before.md` and `-after.md`.++The repository has 161 loadable skills, not the 167 in Β§1: `plugins/ruflo-ai-team/directory/skills/` holds a six-skill copy for the Claude directory submission, which Claude Code does not load from the plugin root. ruflo-ai-team has 6 skills and 6 commands.++**The frontmatter key (Β§2.2).** `disable-model-invocation` is the key Claude Code reads. The 2.1.289 binary validates it in skill frontmatter, and its Skill tool refuses such a skill with: "cannot be used with Skill tool due to disable-model-invocation β¦ Do not replicate this skill's workflow by other means". A live run with `claude -p --model haiku --plugin-dir plugins/ruflo-cost-tracker` that asked the model to call `ruflo-cost-tracker:cost-anomaly` got exactly that refusal, and the model reported the skill absent from its listing. The control run against the unmodified SKILL.md launched the skill. The skill is still registered as a slash command in both runs.++**Which skills (Β§2.2).** 27 skills, each matching one of the ADR's categories and not invoked by the model from anywhere else in `plugins/` (agents, commands, skills, hooks):++- ruflo-cost-tracker, 14 sub-reports that `/ruflo-cost` already reaches: cost-anomaly, cost-benchmark, cost-burn, cost-conversation, cost-counterfactual, cost-diff, cost-export, cost-federation, cost-health, cost-projection, cost-report, cost-session, cost-summary, cost-trend.+- ruflo-metaharness, benchmark, spend-gated or scheduled runs and the one-shot scaffold: harness-bench, harness-evolve, harness-learn, harness-mint, harness-oia-audit, harness-security-bench.+- ruflo-workflows: gaia-submission (publishes to an external leaderboard), gaia-architecture-comparison (report).+- ruflo-core: init-project (one-shot setup), ruflo-status (dashboard).+- ruflo-ruvector: vector-setup. ruflo-federation: federation-init. ruflo-agntcy: agntcy-status.++Kept model-invoked on purpose:+- `cost-budget-check`: `ruflo-agent`'s nested-queen agent runs it as a gate before starting a swarm.+- `music-connect`: `/music connect` tells the model to follow that skill.+- The cost-tracker entry points (`cost-advise`, `cost-ledger`, `cost-local`, `cost-openrouter`, `cost-optimize`, `cost-track`, and the `cost-booster-*` and `cost-compact-context` skills): the model is meant to find these from a plain question.++Three places that told the model to use a skill on this list were updated:+- `ruflo-ruvector/commands/vector.md` now asks the user to run `/ruflo-ruvector:vector-setup`.+- The `cost-analyst` agent now points the user at the matching `/ruflo-cost` subcommand.+- The `metaharness-architect` agent names the `/ruflo-metaharness:<skill>` to run.++Each plugin README has a "User-invoked skills (ADR-507)" table. The choosing rule is in `plugins/README.md`. Commands were left alone: no command was moved to user-invoked.++**Descriptions (Β§2.3).** The 20 skill descriptions over 300 characters were cut to one or two sentences with "Use when β¦" (215β277 characters), as single-line YAML scalars. No command description was over 300. `scripts/audit-skill-frontmatter.mjs`, already a CI step in `all-plugins-smoke.yml`, now:+- fails any skill or command description over 300 characters, measured unfolded so a `>` block cannot slip past;+- fails any `disable-model-invocation: true` skill that is not named under a "User-invoked" heading in its plugin README.++Because the backlog is now zero, the ceiling is a hard limit and not a baseline. Before the README tables were added, this check failed on all 28 then-marked skills, which is the negative control. `music-connect` was unmarked afterwards. The ADR-112 audit (`audit-tool-descriptions.mjs`) covers MCP tool descriptions only. It is unchanged and passes: 414/414.++Three existing descriptions are not valid strict YAML because they contain an unquoted `: `: `cost-openrouter`, `harness-threat-model` and `llm-config`. They are under 300 characters and outside this change, so they were left as they are.++**`skillTrim` (Β§2.4) is not built.** The engine has no hook that hides a skill from the model's listing. Evidence from the ruflo-mods engine types (`.claude-plugin/types/claude-code/index.d.ts`, Claude Code 2.1.289):+- `agent.offer` exists for agents only.+- `skill.prompt` fires when a skill's prompt is expanded, so the skill is already chosen by then. It can only rewrite the text.+- `command.describe`'s `isHidden` hides a command from the typeahead and `/help`, not from the model's skill listing, so it would not save these tokens.+- `ContextSkills`/`ContextSkill` are read-only measurements.++No `skillTrim` or `skillTrimKeep` option was added. This section is closed until the engine gains a skill-listing hook.+diff --git a/v3/docs/validation/plugin-footprint-2026-10-29-after.md b/v3/docs/validation/plugin-footprint-2026-10-29-after.mdnew file mode 100644index 0000000000..16032204aa--- /dev/null+++ b/v3/docs/validation/plugin-footprint-2026-10-29-after.md@@ -0,0 +1,103 @@+# Plugin footprint (after), 2026-10-29++Measured by `node scripts/measure-plugin-footprint.mjs` (ADR-507 Β§2.1) at `921a9ca33`.++The listing is what Claude Code keeps in context on every turn: the name and description of each skill, command and agent the model can invoke. A skill or command with `disable-model-invocation: true` stays runnable as `/<name>` but leaves that listing, so it is not counted. Tokens are estimated as characters Γ· 4.++## Totals++| | Count |+|---|---:|+| Skills | 161 (27 user-invoked only) |+| Commands | 62 (0 user-invoked only) |+| Agents | 65 |+| Listing characters | 39,402 (skills 22,640, commands 5,872, agents 10,890) |+| Estimated tokens per turn | 9,851 |+| Skill/command descriptions over 300 characters | 0 |++These totals assume every ruflo plugin is enabled at once; a session pays only for the plugins it enables.++## Per plugin++| Plugin | Skills (user-invoked) | Commands (user-invoked) | Agents | Listing chars | Est. tokens |+|---|---:|---:|---:|---:|---:|+| ruflo-ai-team | 6 (0) | 6 (0) | 4 | 3,015 | 754 |+| ruflo-neural-trader | 9 (0) | 1 (0) | 4 | 2,760 | 690 |+| ruflo-agent | 5 (0) | 2 (0) | 9 | 2,581 | 645 |+| ruflo-metaharness | 13 (6) | 1 (0) | 1 | 2,180 | 545 |+| ruflo-cost-tracker | 24 (14) | 1 (0) | 1 | 2,168 | 542 |+| ruflo-browser | 10 (0) | 1 (0) | 1 | 1,957 | 489 |+| ruflo-workflows | 5 (2) | 8 (0) | 3 | 1,738 | 435 |+| ruflo-goals | 5 (0) | 1 (0) | 4 | 1,570 | 393 |+| ruflo-core | 5 (2) | 2 (0) | 4 | 1,189 | 297 |+| ruflo-music | 7 (0) | 1 (0) | 2 | 1,103 | 276 |+| ruflo-ddd | 3 (0) | 1 (0) | 1 | 1,046 | 262 |+| ruflo-deepseek-harness | 2 (0) | 1 (0) | 1 | 1,027 | 257 |+| ruflo-iot-cognitum | 5 (0) | 1 (0) | 4 | 990 | 248 |+| ruflo-intelligence | 3 (0) | 2 (0) | 1 | 952 | 238 |+| ruflo-ruvector | 4 (1) | 1 (0) | 1 | 932 | 233 |+| ruflo-adr | 5 (0) | 1 (0) | 1 | 874 | 219 |+| ruflo-ruos | 1 (0) | 4 (0) | 1 | 836 | 209 |+| ruflo-sparc | 3 (0) | 1 (0) | 1 | 816 | 204 |+| ruflo-rure | 1 (0) | 1 (0) | 1 | 772 | 193 |+| ruflo-testgen | 3 (0) | 1 (0) | 1 | 716 | 179 |+| ruflo-ruvllm | 2 (0) | 1 (0) | 1 | 682 | 171 |+| ruflo-aidefence | 2 (0) | 1 (0) | 1 | 677 | 169 |+| ruflo-swarm | 2 (0) | 2 (0) | 2 | 672 | 168 |+| ruflo-agentdb | 2 (0) | 2 (0) | 1 | 638 | 160 |+| ruflo-security-audit | 2 (0) | 1 (0) | 1 | 608 | 152 |+| ruflo-plugin-creator | 3 (0) | 1 (0) | 1 | 607 | 152 |+| ruflo-rag-memory | 2 (0) | 2 (0) | 1 | 565 | 141 |+| ruflo-market-data | 2 (0) | 1 (0) | 1 | 562 | 141 |+| ruflo-jujutsu | 2 (0) | 1 (0) | 1 | 475 | 119 |+| ruflo-federation | 3 (1) | 1 (0) | 1 | 452 | 113 |+| ruflo-docs | 2 (0) | 1 (0) | 1 | 447 | 112 |+| ruflo-observability | 2 (0) | 1 (0) | 1 | 447 | 112 |+| ruflo-loop-workers | 2 (0) | 2 (0) | 1 | 435 | 109 |+| ruflo-autopilot | 2 (0) | 2 (0) | 1 | 429 | 107 |+| ruflo-migrations | 2 (0) | 1 (0) | 1 | 426 | 107 |+| ruflo-knowledge-graph | 2 (0) | 1 (0) | 1 | 390 | 98 |+| ruflo-daa | 2 (0) | 1 (0) | 1 | 385 | 96 |+| ruflo-rvf | 2 (0) | 1 (0) | 1 | 383 | 96 |+| ruflo-bbs-federation | 1 (0) | 0 (0) | 0 | 268 | 67 |+| ruflo-business-pods | 1 (0) | 0 (0) | 0 | 255 | 64 |+| ruflo-console | 1 (0) | 0 (0) | 0 | 247 | 62 |+| ruflo-arena | 0 (0) | 1 (0) | 0 | 130 | 33 |+| ruflo-agntcy | 1 (1) | 0 (0) | 0 | 0 | 0 |++## Descriptions over 300 characters++None.++## User-invoked only (`disable-model-invocation: true`)++| Plugin | Kind | Name |+|---|---|---|+| ruflo-agntcy | skill | agntcy-status |+| ruflo-core | skill | init-project |+| ruflo-core | skill | ruflo-status |+| ruflo-cost-tracker | skill | cost-anomaly |+| ruflo-cost-tracker | skill | cost-benchmark |+| ruflo-cost-tracker | skill | cost-burn |+| ruflo-cost-tracker | skill | cost-conversation |+| ruflo-cost-tracker | skill | cost-counterfactual |+| ruflo-cost-tracker | skill | cost-diff |+| ruflo-cost-tracker | skill | cost-export |+| ruflo-cost-tracker | skill | cost-federation |+| ruflo-cost-tracker | skill | cost-health |+| ruflo-cost-tracker | skill | cost-projection |+| ruflo-cost-tracker | skill | cost-report |+| ruflo-cost-tracker | skill | cost-session |+| ruflo-cost-tracker | skill | cost-summary |+| ruflo-cost-tracker | skill | cost-trend |+| ruflo-federation | skill | federation-init |+| ruflo-metaharness | skill | harness-bench |+| ruflo-metaharness | skill | harness-evolve |+| ruflo-metaharness | skill | harness-learn |+| ruflo-metaharness | skill | harness-mint |+| ruflo-metaharness | skill | harness-oia-audit |+| ruflo-metaharness | skill | harness-security-bench |+| ruflo-ruvector | skill | vector-setup |+| ruflo-workflows | skill | gaia-architecture-comparison |+| ruflo-workflows | skill | gaia-submission |+diff --git a/v3/docs/validation/plugin-footprint-2026-10-29-before.md b/v3/docs/validation/plugin-footprint-2026-10-29-before.mdnew file mode 100644index 0000000000..6782a68ea0--- /dev/null+++ b/v3/docs/validation/plugin-footprint-2026-10-29-before.md@@ -0,0 +1,96 @@+# Plugin footprint (before), 2026-10-29++Measured by `node scripts/measure-plugin-footprint.mjs` (ADR-507 Β§2.1) at `00c0f743f`.++The listing is what Claude Code keeps in context on every turn: the name and description of each skill, command and agent the model can invoke. A skill or command with `disable-model-invocation: true` stays runnable as `/<name>` but leaves that listing, so it is not counted. Tokens are estimated as characters Γ· 4.++## Totals++| | Count |+|---|---:|+| Skills | 161 (0 user-invoked only) |+| Commands | 62 (0 user-invoked only) |+| Agents | 65 |+| Listing characters | 47,743 (skills 30,981, commands 5,872, agents 10,890) |+| Estimated tokens per turn | 11,936 |+| Skill/command descriptions over 300 characters | 20 |++These totals assume every ruflo plugin is enabled at once; a session pays only for the plugins it enables.++## Per plugin++| Plugin | Skills (user-invoked) | Commands (user-invoked) | Agents | Listing chars | Est. tokens |+|---|---:|---:|---:|---:|---:|+| ruflo-cost-tracker | 24 (0) | 1 (0) | 1 | 5,084 | 1,271 |+| ruflo-metaharness | 13 (0) | 1 (0) | 1 | 5,021 | 1,255 |+| ruflo-ai-team | 6 (0) | 6 (0) | 4 | 3,015 | 754 |+| ruflo-neural-trader | 9 (0) | 1 (0) | 4 | 2,760 | 690 |+| ruflo-agent | 5 (0) | 2 (0) | 9 | 2,581 | 645 |+| ruflo-workflows | 5 (0) | 8 (0) | 3 | 2,019 | 505 |+| ruflo-browser | 10 (0) | 1 (0) | 1 | 1,957 | 489 |+| ruflo-core | 5 (0) | 2 (0) | 4 | 1,621 | 405 |+| ruflo-goals | 5 (0) | 1 (0) | 4 | 1,570 | 393 |+| ruflo-deepseek-harness | 2 (0) | 1 (0) | 1 | 1,385 | 346 |+| ruflo-music | 7 (0) | 1 (0) | 2 | 1,103 | 276 |+| ruflo-ruvector | 4 (0) | 1 (0) | 1 | 1,078 | 270 |+| ruflo-ddd | 3 (0) | 1 (0) | 1 | 1,046 | 262 |+| ruflo-iot-cognitum | 5 (0) | 1 (0) | 4 | 990 | 248 |+| ruflo-intelligence | 3 (0) | 2 (0) | 1 | 952 | 238 |+| ruflo-ruos | 1 (0) | 4 (0) | 1 | 906 | 227 |+| ruflo-rure | 1 (0) | 1 (0) | 1 | 903 | 226 |+| ruflo-adr | 5 (0) | 1 (0) | 1 | 874 | 219 |+| ruflo-testgen | 3 (0) | 1 (0) | 1 | 864 | 216 |+| ruflo-sparc | 3 (0) | 1 (0) | 1 | 816 | 204 |+| ruflo-plugin-creator | 3 (0) | 1 (0) | 1 | 742 | 186 |+| ruflo-ruvllm | 2 (0) | 1 (0) | 1 | 682 | 171 |+| ruflo-aidefence | 2 (0) | 1 (0) | 1 | 677 | 169 |+| ruflo-swarm | 2 (0) | 2 (0) | 2 | 672 | 168 |+| ruflo-agentdb | 2 (0) | 2 (0) | 1 | 638 | 160 |+| ruflo-security-audit | 2 (0) | 1 (0) | 1 | 608 | 152 |+| ruflo-business-pods | 1 (0) | 0 (0) | 0 | 581 | 145 |+| ruflo-rag-memory | 2 (0) | 2 (0) | 1 | 565 | 141 |+| ruflo-market-data | 2 (0) | 1 (0) | 1 | 562 | 141 |+| ruflo-federation | 3 (0) | 1 (0) | 1 | 540 | 135 |+| ruflo-jujutsu | 2 (0) | 1 (0) | 1 | 475 | 119 |+| ruflo-docs | 2 (0) | 1 (0) | 1 | 447 | 112 |+| ruflo-observability | 2 (0) | 1 (0) | 1 | 447 | 112 |+| ruflo-loop-workers | 2 (0) | 2 (0) | 1 | 435 | 109 |+| ruflo-autopilot | 2 (0) | 2 (0) | 1 | 429 | 107 |+| ruflo-migrations | 2 (0) | 1 (0) | 1 | 426 | 107 |+| ruflo-knowledge-graph | 2 (0) | 1 (0) | 1 | 390 | 98 |+| ruflo-daa | 2 (0) | 1 (0) | 1 | 385 | 96 |+| ruflo-rvf | 2 (0) | 1 (0) | 1 | 383 | 96 |+| ruflo-agntcy | 1 (0) | 0 (0) | 0 | 376 | 94 |+| ruflo-bbs-federation | 1 (0) | 0 (0) | 0 | 361 | 90 |+| ruflo-console | 1 (0) | 0 (0) | 0 | 247 | 62 |+| ruflo-arena | 0 (0) | 1 (0) | 0 | 130 | 33 |++## Descriptions over 300 characters++| Plugin | Kind | Name | Characters |+|---|---|---|---:|+| ruflo-business-pods | skill | pod-sales | 572 |+| ruflo-metaharness | skill | harness-gepa | 461 |+| ruflo-metaharness | skill | harness-security-bench | 460 |+| ruflo-deepseek-harness | skill | deepseek-reason | 457 |+| ruflo-metaharness | skill | harness-evolve | 441 |+| ruflo-rure | skill | analyze-app | 408 |+| ruflo-deepseek-harness | skill | deepseek-chat | 407 |+| ruflo-metaharness | skill | harness-learn | 406 |+| ruflo-metaharness | skill | harness-similarity | 400 |+| ruflo-metaharness | skill | harness-bench | 395 |+| ruflo-testgen | skill | tdd-repair | 389 |+| ruflo-agntcy | skill | agntcy-status | 363 |+| ruflo-plugin-creator | skill | create-mod | 350 |+| ruflo-metaharness | skill | harness-drift-from-history | 343 |+| ruflo-bbs-federation | skill | cross-host-federation | 340 |+| ruflo-cost-tracker | skill | cost-anomaly | 318 |+| ruflo-metaharness | skill | harness-score | 311 |+| ruflo-metaharness | skill | harness-mint | 310 |+| ruflo-core | skill | witness | 306 |+| ruflo-ruos | skill | ruos-host-run | 306 |++## User-invoked only (`disable-model-invocation: true`)++None.+