Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/phungkaizen/maintainer-autopilot/agents-mdgit clone --depth 1 https://github.com/phungkaizen/maintainer-autopilotWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/phungkaizen/maintainer-autopilot/agents-md)<a href="https://agentmods.dev/instructions/phungkaizen/maintainer-autopilot/agents-md"><img src="https://agentmods.dev/badge/instructions/phungkaizen/maintainer-autopilot/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00320 | $0.00320 |
| Opus 5 | $0.00160 | $0.00160 |
| Sonnet 5 | $0.00064 | $0.00064 |
| Haiku 4.5 | $0.00032 | $0.00032 |
Grade A, and why
maintainer-autopilot AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
AGENTS.md
Project intent
Maintainer Autopilot is a local-first orchestration layer for resumable, auditable AI maintenance pipelines. It is not a general-purpose autonomous shell and must preserve repository safety ahead of convenience.
Required invariants
- Preserve the single-writer invariant. Never create a second active writer while a writer lock exists.
- Never bypass state-machine transitions to make a run appear successful.
- Implementation, deterministic verification, promotion, PR/CI, merge, and checkpoint are separate stages.
- Repair an existing task lineage instead of silently creating a duplicate task or PR.
- Record candidate file hashes before promotion.
- Keep automatic merge disabled by default.
- Do not weaken or delete tests merely to make a gate pass.
- Do not add secrets, tokens, local auth files, or private product data to the repository.
Toolchain
- Node.js 22.
- Zero runtime dependencies is an intentional design constraint for v0.x unless a change clearly justifies otherwise.
- Run
npm ciafter lockfile changes. - Run
npm run checkbefore declaring a candidate ready.
Change discipline
Prefer small, reviewable changes with deterministic tests. New state transitions require tests for both the allowed path and at least one rejected path. New adapters must fail closed when required external tooling or credentials are unavailable.
Codex integration
The default implementation adapter uses codex exec --ephemeral --sandbox workspace-write. Read-only review uses the read-only sandbox. Do not replace these with broader permissions unless the security model and documentation are updated deliberately.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 32 lines · 320 tokens per session scan A 0e3d0937d284
maintainer-autopilot AGENTS.md is an instructions file published in the GitHub repository phungkaizen/maintainer-autopilot (100 stars, last pushed 24d ago), licensed MIT. It adds 320 tokens to every session, about $0.0016 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
vercel-action CLAUDE.md
Instructions for amondnet/vercel-action, covering claude.md, project overview, package manager, essential commands and development.
setup-chromedriver CLAUDE.md
Claude Code instructions for nanasess/setup-chromedriver, covering claude.md, commands, build and development, running a single test and or.
databricks-template CLAUDE.md
Instructions for andre-salvati/databricks-template, covering claude.md, project overview, tooling: mcp servers, cli, skills → see specs/tooling.md, commands and architecture & data model → see specs/.
ConsultMe CLAUDE.md
Claude Code instructions for Tarek-Bohdima/ConsultMe, covering claude.md, project context, common commands, module graph and conventions enforced by tooling.
omen CLAUDE.md
Claude Code instructions for panbanda/omen, covering claude.md, build and development commands, setup git hooks (run once after clone), format code and run linter.
ai-research-arm CLAUDE.md
Claude Code instructions for guzus/ai-research-arm, covering claude.md, project overview, core pipeline architecture, key components and scripts (scripts/).