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/nytc69/review-loop/claude-mdgit clone --depth 1 https://github.com/NYTC69/review-loopWrote 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/nytc69/review-loop/claude-md)<a href="https://agentmods.dev/instructions/nytc69/review-loop/claude-md"><img src="https://agentmods.dev/badge/instructions/nytc69/review-loop/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 | $0.03205 | $0.03205 |
| Opus 5 | $0.01603 | $0.01603 |
| Sonnet 5 | $0.00641 | $0.00641 |
| Haiku 4.5 | $0.00320 | $0.00320 |
Grade C, and why
review-loop CLAUDE.md scanned grade C with 2 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 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.
Strips warnings and disclaimersmediumAnti-refusal
Omitting safety caveats hides risk from the user and is a common jailbreak preamble.
review-loop is designed for a broad audience — not every user will have the same tools installed. Any integration with external tools (MemPalace, Graphify, etc.) **must be strictly optional**: if the tool is unavailable, Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
`~/.codex/config.toml`. How it starts
The opening of the file, as written. The whole thing — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
review-loop — Development Notes
Known Pitfalls
Plugin agent type sandbox bug (CRITICAL)
ALL plugin-defined agent types have their tools silently blocked by Claude Code sandbox. This applies to both tools: all AND tools: read-only. Agents invoked via subagent_type: review-loop:<name> get zero tool access — they cannot Read, Grep, or Bash. The result is tool_uses: 0 and completely hallucinated output.
Fix: Always use subagent_type: general-purpose and inline the agent's .md body in the prompt. For read-only agents, add "Report only, do not modify files" to the prompt.
History: First discovered with Executor (tools: all) in commit 8506809. We incorrectly concluded tools: read-only was safe. This wrong assumption was recorded in memory and carried forward through 4+ commits (including a code-simplifier recurrence on 2026-04-06) until rust-reviewer was caught hallucinating (issue #3). The root cause is Claude Code's sandbox, not the tools: declaration.
Rule: When adding ANY new agent invocation, always use subagent_type: general-purpose with inlined body. Never use subagent_type: review-loop:<name>.
README.md must stay intact (lint SSOT dependency)
run-skill-lint's guide:readme_marks_* and shared-schema:* assertions treat several phrases inside README.md as the single source of truth (SSOT). Trimming or rewriting README content these assertions reach makes lint FAIL (5 cases observed during compass adopt Round 1). The compass-adopt migration handled this by double-storing the migrated blocks: ARCHITECTURE.md / CLAUDE.md / DESIGN.md gained ## Migrated — README.md:<lines> blocks, but the README body was reverted to its full 370-line form so existing lint needles still resolve.
Rule: do not trim or restructure README.md without first updating both the guide skill and the lint contract to point their needles at the new SSOT (e.g. the migrated blocks in CLAUDE.md). The ## Migrated — blocks are intentional duplication, not a cleanup target.
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 · 231 lines · 3,205 tokens per session scan C 2cbced0b256e
review-loop CLAUDE.md is an instructions file published in the GitHub repository NYTC69/review-loop (3 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 3,205 tokens to every session, about $0.0160 per session on Opus 5. A static security scan graded it C with 2 findings (strips warnings and disclaimers, 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
stockbit-mcp CLAUDE.md
Claude Code instructions for INo-xious/stockbit-mcp, covering claude.md, what this is, commands, the map and three invariants. do not break them; each has a test.
agent-plugins-skills copilot-instructions.md
Copilot instructions for richfrem/agent-plugins-skills, covering copilot instructions for agent-plugins-skills, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
AmbyKit CLAUDE.md
Instructions for ambystechcom/AmbyKit, covering claude code — notes for the ambykit repo and claude-specific.
lumen AGENTS.md
AGENTS.md instructions for ory/lumen, covering agents.md, repository expectations and commands.
ai-auto-work CLAUDE.md
Claude Code instructions for chaohong-ai/ai-auto-work, covering gamemaker - ai native 游戏创作平台, 项目简介, ai 工作目录约定 and 按需加载（不默认装载）.
catalyst AGENTS.md
AGENTS.md instructions for coalesce-labs/catalyst, covering agents.md, what this repository is, agent philosophy, build & test and key principles.