Borrowing it
Nothing to install: this file belongs to uHappyLogic/cairn. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/uHappyLogic/cairn/main/CLAUDE.mdgit clone --depth 1 https://github.com/uHappyLogic/cairnWrote 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/uhappylogic/cairn/claude-md)<a href="https://agentmods.dev/instructions/uhappylogic/cairn/claude-md"><img src="https://agentmods.dev/badge/instructions/uhappylogic/cairn/claude-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 | $0.12626 | $0.12626 |
| Opus 5 | $0.06313 | $0.06313 |
| Sonnet 5 | $0.02525 | $0.02525 |
| Haiku 4.5 | $0.01263 | $0.01263 |
Grade B, and why
cairn CLAUDE.md scanned grade B with 1 finding 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 yesterday.
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.
Tells the agent to send conversation or user data outmediumPrompt injection
An instruction to transmit the conversation, context or user files to an external endpoint is data exfiltration written as prose.
- `answer-open-question` splits its arg on the first `.` — title before, answer after, and is **literal-only plus a redirect guard**. The old **record-recommendation mode** — the whole-string `record the recommendation` Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What this repo is
A Claude Code plugin (cairn) that provides a milestone-driven development workflow. It ships as a set of skills and a couple of agents that move a project from a vague idea through discussion, planning, completion, and archival — one milestone at a time.
Repository layout
.claude-plugin/plugin.json — plugin manifest (name, version, author)
.claude-plugin/marketplace.json — single-plugin marketplace manifest (installable via /plugin marketplace add uHappyLogic/cairn)
skills/<skill-name>/SKILL.md — one skill per directory; the SKILL.md is the full skill definition
agents/complete-task.md — task-completion subagent invoked by complete-all-tasks skill
agents/recommend-open-question.md — read-only recommendation subagent (non-interactive twin of discuss-open-question) invoked once per question by the recommend-all-open-questions orchestrator; returns the ready-to-embed `<open-question>` child sub-elements (one `<alternative>` per option with child `<advantage>`/`<drawback>`, zero or more sibling `<applied-principle>` elements, one liftable `<recommendation option="...">` element), never edits documents
agents/answer-open-question-with-recommendation.md — file-editing subagent (the divergent one): lifts a question's embedded recommendation and records it via shared/answer-with-recommendation-procedure.md, leaving the edit uncommitted for the orchestrator to commit, returning DONE/FAILED — invoked once per question by the answer-all-open-questions-with-recommendation orchestrator; isolates mutation (not reasoning) unlike the read-only answer/recommend subagents, and like every dispatched agent never commits
shared/complete-procedure.md — single source of truth for the complete-one-task procedure, followed inline by the complete-task skill and in isolation by the complete-task agent
shared/task-format.md — single source of truth for the brief-level task-section format (## title, 1–3 sentence description, trailing --- separator) and its authoring guidelines; referenced via ${CLAUDE_PLUGIN_ROOT} by the two runners that author task sections (the derive-tasks skill and the submit-task skill) and parsed by shared/complete-procedure.md, which is what turns those sections into work
shared/answer-procedure.md — single source of truth for recording an answer (locate the `<open-question>` block by Short Title via the boundary-line CLI → remove the whole `<open-question …>`…`</open-question>` block → fold decision into ## Decisions → cascade), referenced by answer-open-question and shared/answer-with-recommendation-procedure.md (which composes over it with a lifted ANSWER)
shared/answer-with-recommendation-procedure.md — single source of truth for the lift-then-delegate procedure (locate the `<open-question>` block via the CLI → lift its `<recommendation option="...">` element as "`<option>` — `<rationale>`" with reverse entity-substitution, with a no-`<recommendation>`-element guard → delegate to shared/answer-procedure.md with the derived ANSWER), composing over the execution-neutral recording core; followed inline by the answer-open-question-with-recommendation skill and in isolation by the answer-open-question-with-recommendation agent
shared/recommend-procedure.md — single source of truth for the principle-aware "alternatives + single recommendation for one question" analytical core (execution-neutral, reads the fixed-path principle store in place, no XML rendering markup), referenced by the discuss-open-question skill and the recommend-open-question agent
shared/commit-procedure.md — single source of truth for the skill-layer commit step (dirty-own-path no-op guard → path-scoped git add of the caller's own PATHS → git commit under the caller's function-derived <Marker>: <descriptor> SUBJECT), referenced via ${CLAUDE_PLUGIN_ROOT} by every committing skill and every committing orchestrator; execution-neutral about resolving those paths/subjects (the wrappers own that)
shared/get-current-milestone.md — single source of truth for resolving <MILESTONE_DIR> from the grep-able pointer line in milestones/README.md
milestones/answer_decision_principles.md — project-wide answering-principle store at the milestones/ root, shared across all milestones (above any <MILESTONE_DIR>); the loop's one fixed-path artifact
LICENSE — MIT license
README.md — authoritative workflow documentation and skill reference
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.
- yesterday Changed · +1 lines · +718 tokens per session 49d169393437
- 2d ago Changed · -12 lines · +1,262 tokens per session 946836175976
- 6d ago First seen · 131 lines · 10,646 tokens per session scan B 7b2eed5aaf3e
cairn CLAUDE.md is an instructions file published in the GitHub repository uHappyLogic/cairn (2 stars, last pushed 2d ago), licensed MIT. It adds 12,626 tokens to every session, about $0.0631 per session on Opus 5. A static security scan graded it B with 1 finding (tells the agent to send conversation or user data out). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.