credence-pi — OpenClaw plugin body

The body that lets the credence-pi Bayesian brain govern a real pi/OpenClaw agent. It registers an OpenClaw before_tool_call hook, forwards each proposed tool call to the credence-pi daemon (the opaque brain), and maps the brain's decision back to OpenClaw:

| brain effector | OpenClaw result | |---|---| | proceed | allow | | block | { block: true, blockReason } | | ask | { requireApproval } — OpenClaw's native approval dialog; the user's choice is posted back so the brain learns |

It also logs tool outcomes (after_tool_call) and reconstructs per-turn cost (llm_output token counts × a price table) so the observation log accumulates the data the dollars-saved surface needs.

Fail-open: if the daemon is unreachable or slow, the tool proceeds (one warning per outage). Governance never blocks the agent on infrastructure failure.

This is one of two bodies over the same brain; the other, apps/credence-pi/extension/, targets the pi coding agent directly. Brain + wire (POST /sensor, SSE /signals) are shared and unchanged.

Why a plugin (not a pi extension)

Current OpenClaw vendors pi's coding-agent and runs its gateway agent with noExtensions: true, so a pi ExtensionFactory never loads. The supported interception point is an OpenClaw plugin before_tool_call hook. See docs/credence-pi-pass-2/move-1-design.md.

Install (operator)

1. Start the brain daemon — it listens on http://127.0.0.1:8787. Either run the published image (docker run -p 8787:8787 -v ~/.credence-pi:/root/.credence-pi ghcr.io/gfrmin/credence-pi-daemon) or run it from source (julia --project=<repo-root> apps/credence-pi/daemon/main.jl). See apps/credence-pi/daemon/README.md.
2. Build the plugin: cd apps/credence-pi/openclaw-plugin && npm install && npm run build.
3. Install it into OpenClaw. From a published registry: openclaw plugins install @gfrmin/credence-pi-openclaw; or link a local checkout for development: openclaw plugins install -l apps/credence-pi/openclaw-plugin. Then openclaw plugins enable credence-pi.
4. Per-turn cost signal. On current OpenClaw (≥ 2026.6.2) the llm_output cost hook is active out of the box — no extra config. (Older builds gated it behind a since-removed plugins.entries.credence-pi.hooks.allowConversationAccess flag; that key is now rejected by the config schema.) Governance — allow/block/ask — never depended on it.
5. Restart the gateway so it picks up the plugin.
6. Verify it loaded: openclaw plugins list shows credence-pi as loaded.

Config (openclaw.plugin.json → configSchema)

| key | default | meaning | |---|---|---| | daemonUrl | http://127.0.0.1:8787 | credence-pi daemon base URL | | hookTimeoutMs | 3000 | max wait for the daemon decision before failing open | | approvalTimeoutMs | 120000 | how long OpenClaw waits for the user on an ask before denying | | redactToolInputs | false | omit tool-call inputs from sensor events (they can carry secrets); ask-preview becomes generic | | silent | false | suppress info/warn logs | | pricing | — | per-model USD/Mtok overrides: { "<model>": { "input": n, "output": n, "cacheRead": n, "cacheWrite": n } } |

Privacy: by default the daemon logs proposed_call.input (commands, paths) to ~/.credence-pi/observations.jsonl. Set redactToolInputs: true to keep inputs out of the log. Fail-open warnings are emitted once per outage (re-armed when the daemon recovers).

The built-in price table is approximate; set pricing for an exact dollars-saved figure for your providers. Unknown/unpriced models log usd: null (token counts still recorded; the Move-2 surface applies a token×price fallback).

Develop

npm install
npm run build      # tsc → dist/
npm run typecheck  # tsc --noEmit
npm test           # node --test (tsx) over tests/*.test.ts

Known limitations (Move 1 / MVP-0)

• working-directory-relative and time-since-last-user-message features are best-effort (OpenClaw doesn't put cwd or message timestamps on the tool ctx); the loop-relevant features (tool-name, parent, repetition) are exact. In Move 1 the brain does not condition on features at decision time, so this only affects future feature-conditioned learning.
• Cost USD is reconstructed from a local price table (no host calculateCost dependency); override via pricing.
• The per-run feature buffer is bounded per run but the run map is not evicted on a long-lived gateway — a PASS-2 cleanup item.