Skip to content

armgit / ruflo

Public Imported from GitHub
0

🌊 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

feat(cost-tracker): P11 β€” cost-per-conversation view (v0.12.0)

Reuven088e18f174b9
diff --git a/plugins/ruflo-cost-tracker/.claude-plugin/plugin.json b/plugins/ruflo-cost-tracker/.claude-plugin/plugin.json
index b2f56e1e76..d1dc172dbc 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": "0.11.0",
+ "version": "0.12.0",
"author": {
"name": "ruvnet",
"url": "https://github.com/ruvnet"
diff --git a/plugins/ruflo-cost-tracker/commands/ruflo-cost.md b/plugins/ruflo-cost-tracker/commands/ruflo-cost.md
index 1043cd91cc..79c85934a8 100644
--- a/plugins/ruflo-cost-tracker/commands/ruflo-cost.md
+++ b/plugins/ruflo-cost-tracker/commands/ruflo-cost.md
@@ -55,6 +55,11 @@ Cost tracking commands:
3. The script wraps `npx @claude-flow/cli hooks model-outcome -t ... -m ... -o ...` with explicit-argv spawnSync so quoting is safe
4. Without this, the router doesn't learn from cost-optimize recommendations and the Tier 1 bypass rate doesn't tighten over time
+**`cost conversation`** -- Per-conversation cost view: list every session in `cost-tracking` with started-at, message count, top model, total cost. Different lens from `cost report` (which is per-agent/per-model).
+1. Run `node plugins/ruflo-cost-tracker/scripts/conversation.mjs`
+2. Optional `CONV_FORMAT=json`, `CONV_LIMIT=N`, `CONV_NAMESPACE=...`
+3. Reports: total across conversations, per-tier rollup, per-session table
+
**`cost trend`** -- Read all docs/benchmarks/runs/*.json and surface drift in the gate metrics β€” win rate, avg latency, p99, escalation rate, speedup vs LLM. Flags regressions the binary smoke gate misses.
1. Run `node plugins/ruflo-cost-tracker/scripts/trend.mjs`
2. Optional `TREND_FORMAT=json` for machine-readable output, `TREND_LIMIT=N` to truncate
diff --git a/plugins/ruflo-cost-tracker/scripts/conversation.mjs b/plugins/ruflo-cost-tracker/scripts/conversation.mjs
new file mode 100755
index 0000000000..576143090f
--- /dev/null
+++ b/plugins/ruflo-cost-tracker/scripts/conversation.mjs
@@ -0,0 +1,100 @@
+#!/usr/bin/env node
+// cost-conversation β€” list all session records in the cost-tracking namespace
+// and show cost-per-conversation. Until now reports were per-agent / per-model;
+// this gives a per-conversation timeline.
+//
+// Usage:
+// node scripts/conversation.mjs # markdown table
+// CONV_FORMAT=json node scripts/conversation.mjs # JSON
+// CONV_LIMIT=20 node scripts/conversation.mjs # most recent N
+// CONV_NAMESPACE=cost-tracking (default)
+
+import { spawnSync } from 'node:child_process';
+
+const NS = process.env.CONV_NAMESPACE || 'cost-tracking';
+
+function memoryListSessionKeys() {
+ const r = spawnSync('npx', [
+ '@claude-flow/cli@latest', 'memory', 'list',
+ '--namespace', NS, '--format', 'json',
+ ], { stdio: ['ignore', 'pipe', 'pipe'], encoding: 'utf-8' });
+ if (r.status !== 0) return [];
+ const m = /\[[\s\S]*\]/.exec(r.stdout || '');
+ if (!m) return [];
+ let entries;
+ try { entries = JSON.parse(m[0]); } catch { return []; }
+ return entries
+ .map((e) => e.key)
+ .filter((k) => typeof k === 'string' && k.startsWith('session-'));
+}
+
+function memoryRetrieve(key) {
+ const r = spawnSync('npx', [
+ '@claude-flow/cli@latest', 'memory', 'retrieve',
+ '--namespace', NS, '--key', key,
+ ], { stdio: ['ignore', 'pipe', 'pipe'], encoding: 'utf-8' });
+ if (r.status !== 0) return null;
+ const m = /\{[\s\S]*\}/.exec(r.stdout || '');
+ if (!m) return null;
+ try { return JSON.parse(m[0]); } catch { return null; }
+}
+
+function fmtUsd(n) { return `$${(n || 0).toFixed(4)}`; }
+function fmtTs(s) { return s ? new Date(s).toISOString().replace('T', ' ').replace(/\.\d+Z/, '') : ''; }
+function topModel(byModel) {
+ const entries = Object.entries(byModel || {}).sort((a, b) => (b[1].cost_usd || 0) - (a[1].cost_usd || 0));
+ return entries[0]?.[0] || 'β€”';
+}
+
+function main() {
+ const keys = memoryListSessionKeys();
+ if (!keys.length) {
+ console.log(`No sessions in '${NS}'. Run \`cost track\` first.`);
+ return;
+ }
+ const records = keys.map((k) => memoryRetrieve(k)).filter(Boolean);
+ records.sort((a, b) => (a.startedAt || '').localeCompare(b.startedAt || ''));
+
+ const limit = parseInt(process.env.CONV_LIMIT || '50', 10);
+ const recent = records.slice(-limit);
+
+ const total = records.reduce((s, r) => s + (r.total_cost_usd || 0), 0);
+ const byTier = { haiku: 0, sonnet: 0, opus: 0, unknown: 0 };
+ for (const r of records) {
+ if (r.byTier) {
+ for (const [t, v] of Object.entries(r.byTier)) byTier[t] = (byTier[t] || 0) + v;
+ }
+ }
+
+ if (process.env.CONV_FORMAT === 'json') {
+ console.log(JSON.stringify({
+ conversationCount: records.length,
+ shownCount: recent.length,
+ total_cost_usd: total,
+ byTier,
+ conversations: recent,
+ }, null, 2));
+ return;
+ }
+
+ console.log(`# cost-per-conversation β€” ${records.length} conversations in '${NS}'`);
+ console.log('');
+ console.log(`| Metric | Value |`);
+ console.log(`|---|---:|`);
+ console.log(`| Conversations tracked | ${records.length} |`);
+ console.log(`| **Total cost (all)** | **${fmtUsd(total)}** |`);
+ for (const t of ['haiku', 'sonnet', 'opus', 'unknown']) {
+ if (byTier[t] > 0) console.log(`| Tier ${t} | ${fmtUsd(byTier[t])} |`);
+ }
+ console.log('');
+ console.log(`## ${recent.length === records.length ? 'All' : 'Most recent ' + recent.length} conversations`);
+ console.log('');
+ console.log('| Started | Session | Messages | Top model | Cost |');
+ console.log('|---|---|---:|---|---:|');
+ for (const r of recent) {
+ const sid = (r.sessionId || '').slice(0, 8);
+ console.log(`| ${fmtTs(r.startedAt)} | \`${sid}\` | ${r.messageCount || 0} | \`${topModel(r.byModel)}\` | ${fmtUsd(r.total_cost_usd)} |`);
+ }
+}
+
+main();
diff --git a/plugins/ruflo-cost-tracker/scripts/smoke.sh b/plugins/ruflo-cost-tracker/scripts/smoke.sh
index 2e7711e50c..0054824082 100755
--- a/plugins/ruflo-cost-tracker/scripts/smoke.sh
+++ b/plugins/ruflo-cost-tracker/scripts/smoke.sh
@@ -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 0.11.0 with new keywords"
+step "1. plugin.json declares 0.12.0 with new keywords"
v=$(grep -E '"version"' "$ROOT/.claude-plugin/plugin.json" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)
-if [[ "$v" != "0.11.0" ]]; then
- bad "expected 0.11.0, got '$v'"
+if [[ "$v" != "0.12.0" ]]; then
+ bad "expected 0.12.0, got '$v'"
else
miss=""
for k in namespace-routing mcp agentic-flow agent-booster tier1-routing model-routing benchmarking verified telemetry budget; do
@@ -20,9 +20,9 @@ else
[[ -z "$miss" ]] && ok || bad "missing keywords:$miss"
fi
-step "2. all nine skills present with valid frontmatter"
+step "2. all ten skills present with valid frontmatter"
miss=""
-for s in cost-report cost-optimize cost-booster-route cost-booster-edit cost-compact-context cost-benchmark cost-track cost-budget-check cost-trend; do
+for s in cost-report cost-optimize cost-booster-route cost-booster-edit cost-compact-context cost-benchmark cost-track cost-budget-check cost-trend cost-conversation; do
f="$ROOT/skills/$s/SKILL.md"
[[ -f "$f" ]] || { miss="$miss missing-$s"; continue; }
for k in 'name:' 'description:' 'allowed-tools:'; do
@@ -321,6 +321,19 @@ step "39. ruflo-cost.md documents 'cost trend' subcommand"
grep -q "cost trend" "$ROOT/commands/ruflo-cost.md" \
&& ok || bad "missing"
+step "40a. cost-conversation skill + conversation.mjs"
+F1="$ROOT/scripts/conversation.mjs"
+F2="$ROOT/skills/cost-conversation/SKILL.md"
+miss=""
+[[ -x "$F1" ]] || miss="$miss conv-not-executable"
+node --check "$F1" 2>/dev/null || miss="$miss syntax-error"
+grep -q "memoryListSessionKeys" "$F1" || miss="$miss no-list-fn"
+grep -q "byTier" "$F1" || miss="$miss no-tier-aggr"
+[[ -f "$F2" ]] || miss="$miss skill-missing"
+grep -q "conversation\.mjs" "$F2" || miss="$miss skill-no-script-ref"
+grep -q '^allowed-tools:[[:space:]]*\*' "$F2" && miss="$miss wildcard"
+[[ -z "$miss" ]] && ok || bad "$miss"
+
step "40. CI workflow (smoke + booster bench on PR) present"
WF="$ROOT/../../.github/workflows/cost-tracker-smoke.yml"
miss=""
diff --git a/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md b/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md
new file mode 100644
index 0000000000..d62fb8f7d2
--- /dev/null
+++ b/plugins/ruflo-cost-tracker/skills/cost-conversation/SKILL.md
@@ -0,0 +1,38 @@
+---
+name: cost-conversation
+description: Per-conversation cost view β€” list every session in cost-tracking with started-at, message count, top model, and total cost
+argument-hint: ""
+allowed-tools: Bash
+---
+
+# Cost per Conversation
+
+`cost-report` and `cost-optimize` aggregate by **agent** and **model**. This skill aggregates by **conversation (session)** β€” a different lens that surfaces *which conversations cost the most*. Useful for retrospectives ("which sessions ran long on Opus?") and for evaluating whether a given project's session pattern is sustainable.
+
+## When to use
+
+- After multiple sessions, to see total spend per conversation.
+- Before scoping a long session, to understand typical cost-per-conversation.
+- For per-project rollups via `CONV_NAMESPACE=cost-tracking-<project>`.
+
+## Steps
+
+1. **Run the script** from anywhere:
+
+ ```bash
+ node plugins/ruflo-cost-tracker/scripts/conversation.mjs
+ ```
+
+ Optional env:
+ - `CONV_FORMAT=json` β€” emit JSON instead of markdown
+ - `CONV_LIMIT=20` β€” show only the most recent N conversations
+ - `CONV_NAMESPACE=cost-tracking` β€” override target namespace
+
+2. **Inspect the markdown table** β€” total cost across all conversations, per-tier rollup, then a per-session table (started-at, sessionId prefix, message count, top model, cost).
+
+## Cross-references
+
+- `cost-track` β€” the producer that populates `cost-tracking:session-*`
+- `cost-report` β€” same data, per-agent / per-model lens
+- `cost-trend` β€” drift across bench runs (different axis: corpus runs vs conversations)
+- `cost-budget-check` β€” sums across conversations to evaluate the budget threshold