Borrowing it
Nothing to install: this file belongs to tatargabor/set-agent-comm. 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/tatargabor/set-agent-comm/master/CLAUDE.mdgit clone --depth 1 https://github.com/tatargabor/set-agent-commWrote 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/tatargabor/set-agent-comm/claude-md)<a href="https://agentmods.dev/instructions/tatargabor/set-agent-comm/claude-md"><img src="https://agentmods.dev/badge/instructions/tatargabor/set-agent-comm/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/tatargabor/set-agent-comm/claude-md"><img src="https://agentmods.dev/badge/instructions/tatargabor/set-agent-comm/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.06676 | $0.06676 |
| Opus 5.5 | $0.02670 | $0.02670 |
| Sonnet 5.5 | $0.01335 | $0.01335 |
| Haiku 4.5 | $0.00668 | $0.00668 |
Grade A, and why
set-agent-comm CLAUDE.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 8d 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 — 358 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.
Commands
npm install # one runtime dep: @modelcontextprotocol/sdk
npm test # node --test test/*.test.mjs + smoke-mcp + demo/smoke + demo/smoke-remote
node --test test/store.test.mjs # one file
node --test --test-name-pattern="lost update" # one test
node test/smoke-mcp.mjs # MCP round trip over a real stdio server
npm run demo:smoke # harness + fake `claude`, free
npm run http # HTTP transport, 127.0.0.1:7510/mcp/<agent>
RELAY_SECRET=… npm run relay # the relay, PORT 7511
npm run demo # LIVE run: real `claude -p` sessions, ~$3 and ~30 min
npm run live # LIVE acceptance: kept-alive `claude -p` sessions (LIVE_MODEL, default haiku), ~1 min
npm run demo / demo:remote spend real money and take half an hour. Use npm run demo:smoke
for anything about the harness itself; the live runs only answer "will six real sessions actually
write that way", which no unit test can.
The relay tests (test/relay.test.mjs, test/security.test.mjs) spawn a real relay process and
talk HTTP to it; they need free ports and a few seconds.
Architecture
One core, several thin faces. Everything that touches state goes through src/store.mjs; nothing
else is allowed to know the on-disk layout.
src/store.mjs the core: registry + channel. ZERO runtime deps, fully SYNCHRONOUS —
hooks and cron import it where there is no node_modules and no event loop.
src/tools.mjs the MCP tool definitions + handlers, transport-independent
src/stdio.mjs transport 1 (DEFAULT): identity = cwd, unforgeable
src/http.mjs transport 2: identity = URL path (/mcp/<agent>) + a per-agent token
(`sac http-token`) — WITHOUT one, any local process may connect as any agent
bin/sac.mjs the CLI on the same core, so CLI and MCP cannot drift apart
hooks/*.mjs session-start (check in + watchPaths) and stop (block on unread mail).
Both go SILENT for a headless `claude -p` — see `store.headless`
src/bridge.mjs the remote leg, client half: push/pull, remote↔local name translation
src/relay.mjs the remote leg, server half: stateless, in-memory, HMAC tokens
src/crypto.mjs AES-256-GCM room encryption + stateless token signing
src/triage.mjs the letterbox: a cheap model deciding whether an entry is worth a turn
src/policy.mjs the authorization layer's evaluator — PURE, no I/O but reading the policy
file. Four verdicts: serve / gate / wake / deny. Nothing calls it yet
src/doorbell.mjs the doorbell's I/O: one line into a session's own inbox socket (async, so it
stays out of the synchronous core, which only decides and records)
hooks/watch-guard.mjs PreToolUse on Monitor: declines an inbox watch where the doorbell works
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.
- 8d ago Changed · +29 lines · +737 tokens per session 3f832610aa7c
- 2mo ago First seen · 329 lines · 5,939 tokens per session scan A 6752a052cee0
set-agent-comm CLAUDE.md is an instructions file published in the GitHub repository tatargabor/set-agent-comm (3 stars, last pushed 8d ago), licensed MIT. It adds 6,676 tokens to every session, about $0.0267 per session on Opus 5.5. 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
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).
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.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
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.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.