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/stanley-projects/insomnia/claude-mdgit clone --depth 1 https://github.com/stanley-projects/InsomniaWhat 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 | $0.02990 | $0.02990 |
| Opus 5 | $0.01495 | $0.01495 |
| Sonnet 5 | $0.00598 | $0.00598 |
| Haiku 4.5 | $0.00299 | $0.00299 |
Grade B, and why
Insomnia 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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- [Claude] Stable base: v1.3.0 shipped and tagged. Core detection logic uses SESSION_TIMEOUT_MS=3min (hook inactivity), SESSION_PROCESS_GRACE_MS=5min (process alive + last hook within window), PROCESS_GRACE_MS=30s (proce How it starts
The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Insomnia
Lightweight Windows app (Electron) that prevents the PC from sleeping while AI coding tools (Claude Code, Cursor, Aider, Codex, Ollama) or watched apps are running. Tray icon + manual toggle. Hook-based detection for Claude Code and Codex; process-based for the rest.
Agent Memory Protocol
Use this file as the standing project memory shared between Claude Code and Codex.
Rules:
- Read this file before starting non-trivial work so the current architecture, release state, and prior regressions are fresh.
- Any meaningful code change, architecture decision, bug fix, regression, release action, or workflow adjustment should be recorded here.
- Each note should be prefixed with the agent name in brackets, for example
[Claude]or[Codex]. - Prefer concise entries that explain what changed, why it changed, and any follow-up risk.
- Keep older entries unless they are clearly obsolete; append instead of rewriting history unless the file needs cleanup.
- If the file's encoding gets damaged, normalize it back to plain ASCII or valid UTF-8 while preserving meaning.
- Keep the
Current Statesection accurate when release status, version, detection behavior, or known limitations change.
Agent Communication Protocol
Use this section as the shared communication contract for Claude Code, Codex, or any other coding agent working in this repo.
Rules:
- Start by checking
git status, reading this memory file, and identifying the active branch before making edits. - Preserve user work. Do not revert unrelated dirty files or generated outputs unless the user explicitly asks.
- Before edits, state what you are about to change and why. While working, share concise progress updates when the task takes more than a quick command.
- When blocked, report the blocker, what was already tried, and the safest next option instead of silently changing direction.
- Before finishing, summarize changed files, verification performed, and any remaining risk or follow-up.
- If a change affects behavior, release workflow, installer output, hooks, or package manager metadata, add an
Agent Logentry.
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 · 80 lines · 2,990 tokens per session scan B 32cb97d049f5
Insomnia CLAUDE.md is an instructions file published in the GitHub repository stanley-projects/Insomnia (2 stars, last pushed 3mo ago), licensed MIT. It adds 2,990 tokens to every session, about $0.0149 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
swob AGENTS.md
Instructions for IvyYang1999/swob, covering swob 研发协作唯一规范, 不可越过的边界, 角色与职责, 生命周期 and 分支、worktree 与提交.
ChatCrystal AGENTS.md
Instructions for ZengLiangYi/ChatCrystal, covering agents.md, project overview, commands, development (server port 3721 + client port 13721) and build for production.
cursor-studio copilot-instructions.md
Instructions for DearLicy/cursor-studio, a project described as: Local-first Windows control center for Cursor: manage AI providers, balances, usage and costs, sessions, prompts, MCP, Skills, appearance, and integrations. | Cursor 本地桌面管理工具.
host-console AGENTS.md
AGENTS.md instructions for tuweihuasheng/host-console: Run the local server yourself and open the preview in the browser available to this environment. Do not give the user server-start instructions when you can run it.
swob CLAUDE.md
Instructions for IvyYang1999/swob, a project described as: Local-first git graph and session debugger for Claude Code, Codex, Cursor, OpenCode, and 7 more AI coding harnesses.
mcp-config-manager CLAUDE.md
Claude Code instructions for itsocialist/mcp-config-manager, covering claude.md, behavioral guidelines, 🤖 context self-management instructions, automatic context updates (claude code should follow these) and current sprint tasks (auto-update 2025-10-06).