Scrub Docs Code Parity
Documentsscrub-docs-code-parityScrub Docs Code Parity is an Agent Skill for documents published in the marin-community/marin repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Scheduled scrub: docs and code parity.
QUICK START
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/marin-community/marin/blob/HEAD/.agents/skills/scrub-docs-code-parity/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/scrub-docs-code-parity, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
scrub-docs-code-parity
Use this skill on scheduled scrub turns for docs/code parity in marin-community/marin.
Focus
- Prioritize high-confidence drift in
README.md,docs/,.agents/docs/, and operator-facing scripts. - Confirm command examples match current tooling conventions (
uv run, repo scripts, and documented workflows). - Apply concrete corrections when drift is real; avoid status-only updates.
Decision Heuristics
- Prefer updating docs when current behavior is clear and intentional.
- If implementation is clearly wrong relative to documented intent, update code and docs together.
- Keep scope small and land one useful parity improvement per run when possible.
- If no material drift is found, choose a no-op outcome and keep cadence near daily.
Output
- Keep the final scrub response concise and action-focused.
- Treat local-only edits as incomplete work. If you modify files, publish the result (commit/push and open or update a PR per
.agents/skills/commit/SKILL.md) before finishing this scrub run. - If publish is blocked (auth, permissions, CI infra, etc.), report the blocker and set a future
needs_followup_atinstead of ending the run. - If no material drift is found, explicitly report inspected scope and why no change was needed.
- End the run with exactly one footer line of valid one-line JSON:
HARNESS_SCRUB_LOOP {"needs_followup_at":null}. Setneeds_followup_atto null when the run is complete, or a future RFC 3339 timestamp when another follow-up turn is needed.