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/shenchangmin/cairnkit/agents-mdgit clone --depth 1 https://github.com/shenchangmin/cairnkitWhat 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 | $0.02251 | $0.02251 |
| Opus 5 | $0.01125 | $0.01125 |
| Sonnet 5 | $0.00450 | $0.00450 |
| Haiku 4.5 | $0.00225 | $0.00225 |
Grade A, and why
cairnkit 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 2d 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.
How it starts
The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cairnkit — Agent instructions (Codex / AGENTS.md harness)
cairnkit is a knowledge-precipitation delivery harness. The workflow is the pipe; knowledge
is the moat. This file is the entry point for harnesses that read AGENTS.md (OpenAI Codex
CLI and compatibles). The Claude Code form of the same harness lives in .claude-plugin/ +
commands/ + skills/; both forms drive the same cairn engine and the same role
content in agents/ — see docs/ADAPTERS.md.
The engine: cairn
The deterministic core is a single self-contained binary, cairn (Rust; no language-runtime
dependency). System Git is invoked only by optional shared-repository operations. The binary owns
the state machine, admission gates, and knowledge layer. You never hold workflow state in your
head — cairn is the only writer of state (.cairnkit/STATE.yaml). All commands run from the
host project root:
cairn --root . config show # is the project initialised? (has_run flag)
cairn --root . state init --run-id <id> # start a run
cairn --root . state new --run-id <id> # archive DONE and start the next run
cairn --root . state show | resume # current state / {stage, paused}
cairn --root . state advance # -> next stage (gate-checked)
cairn --root . state set-path-mode <full|lite|single|tooling>
cairn --root . state approve-clarify | unblock
cairn --root . state fail --stage <BUILD_VERIFY|E2E_VERIFY>
cairn --root . kb build-index | query --stage <S> --budget 300 | extract --from <dir> | touch --from <dir>
cairn --root . lifecycle promote | decay · lint [--fix] · knowledge stats
cairn --root . kbrepo sync-in | sync-out | pull | push --message <m> | promote --id <id> --to <L1|L2> | init
Exit codes: 0 ok · 2 usage · 3 gate refused · 4 STATE corrupt. Never edit
.cairnkit/STATE.yaml by hand.
How role dispatch works here
The installed cairnkit plugin is skills-only: its bundled cairnkit and
workflow-orchestrator skills drive the same cairn engine. Source role content is canonical in
agents/ and is mirrored into the isolated native package under plugins/cairnkit/resources/.
It does not require, install, copy, or select custom Codex profiles.
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.
- 2d ago First seen · 132 lines · 2,251 tokens per session scan A 84925005ff67
cairnkit AGENTS.md is an instructions file published in the GitHub repository shenchangmin/cairnkit (1 stars, last pushed 1mo ago), licensed MIT. It adds 2,251 tokens to every session, about $0.0113 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 2026-08-31.
Other instructions, from other repositories
kaas CLAUDE.md
Instructions for bybit-exchange/kaas, covering claude.md, behavioral guidelines, 1. think before coding, 2. simplicity first and 3. surgical changes.
Zettelgarden CLAUDE.md
Instructions for Zettelgarden/Zettelgarden, covering claude.md, project overview, development commands, frontend (zettelkasten-front/) and backend (go-backend/).
Zettelgarden AGENTS.md
Instructions for Zettelgarden/Zettelgarden, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
data-olympus AGENTS.md
Instructions for knaisoma/data-olympus, covering data olympus, project overview, architecture, development workflow and code standards.
hyalo CLAUDE.md
Instructions for ractive/hyalo, covering agents, documentation, dogfooding, rust and language server.
stigmergy CLAUDE.md
Instructions for sturlese/stigmergy, covering stigmergy project doctrine, invariants, package orientation, working rules and testing.