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

docs(ruflo-swarm): allow-list note for the ruflo-mods refuse-risky trust gate

ruv6bd7b2ff7042
diff --git a/plugins/ruflo-swarm/README.md b/plugins/ruflo-swarm/README.md
index 4a8212f696..3af6bb90ba 100644
--- a/plugins/ruflo-swarm/README.md
+++ b/plugins/ruflo-swarm/README.md
@@ -93,6 +93,10 @@ commands, skills and agents above behave as they always have.
network), `routeThreshold`, `injectSpawnContext` (off by default: appends a one-line swarm note to spawned subagents'
prompts) and `audit` (off by default: event names, tool names and agent ids into ruflo memory, never content).
- The mod does not register `prompt.submit` or `tool.check` hooks, and its `tool.call` hook only observes.
+- **With the ruflo-mods trust gate:** if you set `modTrust: refuse-risky`, the gate refuses this mod. The mod hooks
+ `tool.call` and, with `audit` on, `*`, and it calls `process.run`. To keep it, add its provenance to `modTrustAllow`:
+ `ruflo-swarm@ruflo`. The gate matches name@marketplace, not the plugin name. Under the default `modTrust: observe`
+ the mod loads unchanged.
Tests run on the engine's own kit: `claude plugin test plugins/ruflo-swarm` (run `scripts/fetch-mod-types.sh` once
first if you want `tsc -p plugins/ruflo-swarm` to typecheck). The pure readers also run under vitest: