openclaw-autoresearch

Autonomous experiment loop for any optimization target.

Faithful OpenClaw port of davebcn87/pi-autoresearch.

How it works

The agent runs a loop: edit code, run a benchmark, measure the result, keep or discard. Each iteration is logged. The loop runs autonomously until interrupted.

Three tools drive the loop:

Tool	What it does
init_experiment	Configures the session: name, primary metric, unit, direction (lower/higher). Re-calling starts a new segment.
run_experiment	Executes a shell command, times it, captures stdout/stderr, returns pass/fail via exit code.
log_experiment	Records the result. keep auto-commits to git. discard/crash log without committing. Tracks secondary metrics alongside the primary.

Each tool also accepts an optional cwd so callers can target a nested repo explicitly instead of relying on the current session working directory.

All state lives in four repo-root files:

File	Purpose
autoresearch.md	Session doc: objective, metrics, files in scope, constraints, what's been tried. A fresh agent reads this to resume.
autoresearch.sh	Benchmark script. Outputs METRIC name=number lines.
autoresearch.jsonl	Structured log: config headers + experiment entries (metric, status, timestamp, segment, commit hash).
autoresearch.ideas.md	Backlog of promising ideas not yet tried. Optional.

The design is file-first: any agent can pick up the repo-root files and continue the loop without prior context.

Install

npm install

Then load this repo path in OpenClaw plugin discovery and restart the gateway:

plugins:
  load:
    paths:
      - /absolute/path/to/openclaw-autoresearch
  entries:
    openclaw-autoresearch:
      enabled: true

OpenClaw discovers openclaw.plugin.json, loads extensions/openclaw-autoresearch/index.ts, and exposes autoresearch-create.

Manual install is also possible: copy the plugin root, extensions/openclaw-autoresearch/, and skills/autoresearch-create/ into your managed OpenClaw locations, then restart.

Verify:

• skill: autoresearch-create
• tools: init_experiment, run_experiment, log_experiment
• command: /autoresearch (recommended)
• direct skill fallback: /skill autoresearch-create

Prefer the explicit /autoresearch command surface in OpenClaw. The auto-generated native skill alias /autoresearch_create may not trigger reliably on some hosts, so use /skill autoresearch-create if you need to invoke the skill directly.

Use

In the repo you want to optimize:

1. Load the plugin.
2. Run /autoresearch or /autoresearch setup <goal>.
3. Send a normal message with the goal, command, metric (+ direction), files in scope, and constraints.
4. If you need the raw skill invocation, use /skill autoresearch-create.
5. The agent writes autoresearch.md and autoresearch.sh, runs a baseline, then starts looping.
6. Use /autoresearch or /autoresearch status to re-prime context on a later turn.

To resume an existing session, a new agent reads the repo-root files and continues from where the last one stopped.

User steers

Messages sent while an experiment is running are queued and surfaced after the next log_experiment. The agent finishes the current experiment before incorporating the steer.

Ideas backlog

When the agent discovers promising but complex ideas mid-loop, it appends them to autoresearch.ideas.md. On resume, the agent reads the backlog, prunes stale entries, and uses the remaining ideas as experiment paths.

Upstream reference

This port preserves upstream semantics, names, and file contracts while adapting presentation to OpenClaw. There is no Pi-style widget, dashboard, or editor shortcut layer. Remaining differences are tracked in docs/non-parity.md.

• upstream repo: https://github.com/davebcn87/pi-autoresearch
• pinned upstream commit: 2227029fa5712944a36938b5fe59f709cb30ed22 (2227029f)

Validation

npm install --include=dev
npm run typecheck
npm test
npm run validate

The local test shim supports typechecking and tests without a full OpenClaw host checkout. Runtime behavior depends on a real OpenClaw host.

License

MIT