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/marras0914/cordon/claude-mdgit clone --depth 1 https://github.com/marras0914/cordonWrote 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/marras0914/cordon/claude-md)<a href="https://agentmods.dev/instructions/marras0914/cordon/claude-md"><img src="https://agentmods.dev/badge/instructions/marras0914/cordon/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 | $0.03834 | $0.03834 |
| Opus 5 | $0.01917 | $0.01917 |
| Sonnet 5 | $0.00767 | $0.00767 |
| Haiku 4.5 | $0.00383 | $0.00383 |
Grade C, and why
cordon CLAUDE.md scanned grade C 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 4d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
If a future publish 403's with no obvious cause, run a stub-README publish to confirm; if that succeeds, bisect the README content. Likely culprits: any literal SQL injection demo, classic XSS payload (`<script>alert(1)< How it starts
The opening of the file, as written. The whole thing — 245 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cordon
Security gateway for AI agents. Sits between the LLM client (Claude Desktop, Cursor) and MCP servers, enforcing policies, logging every tool call, and requiring human approval for dangerous operations.
Project Layout
packages/
sdk/ @getcordon/policy — defineConfig() helper + all TypeScript types (no runtime deps)
core/ @getcordon/core — proxy engine: gateway, policy, approvals, audit, upstream manager
cli/ @getcordon/cli — CLI commands: `cordon init`, `start`, `login`, `logout`, `replay`
examples/
security-showcase/ — interactive demo + block-test.ts integration tests (7/7 passing)
Planning docs live in the parent directory (../cordon-deux/) alongside the code.
Commands
npm install # install all workspace deps
npm run build # build all packages via turbo (respects dependency order: sdk → core → cli)
npm run dev # watch mode for all packages
npm test # run unit tests (vitest — 148 tests in @getcordon/core as of 2026-07-28)
Building a single package:
cd packages/core && npm run build
Running tests:
npm test # all unit tests via turbo
cd examples/security-showcase && npx tsx block-test.ts # integration test (5/5)
Running the interactive demo:
cd examples/security-showcase && npm run demo
# When the approval prompt appears, type A to approve or D to deny
Architecture
The proxy is an aggregator: one Cordon process presents a unified MCP server to Claude Desktop and internally manages N child MCP processes (one per configured server).
Claude Desktop ──stdio──▶ CordonGateway ──stdio──▶ [MCP server A]
──stdio──▶ [MCP server B]
Critical: process.stdin and process.stdout are owned by the MCP transport. All logging and approval UI must write to process.stderr. The terminal approval channel reads from /dev/tty (Unix) or \\.\CONIN$ (Windows) directly — NOT from stdin.
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.
- 4d ago First seen · 245 lines · 3,834 tokens per session scan C 791934800a41
cordon CLAUDE.md is an instructions file published in the GitHub repository marras0914/cordon (1 stars, last pushed 1mo ago), licensed MIT. It adds 3,834 tokens to every session, about $0.0192 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). 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).
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.
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 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.
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-release stance: foundation over blast radius, repository layout, commands and host sandbox failures.