Borrowing it
Nothing to install: this file belongs to NYTC69/review-loop. 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/NYTC69/review-loop/main/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/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/nytc69/review-loop/claude-md"><img src="https://agentmods.dev/badge/instructions/nytc69/review-loop/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 | $0.02470 | $0.02470 |
| Opus 5.5 | $0.00988 | $0.00988 |
| Sonnet 5.5 | $0.00494 | $0.00494 |
| Haiku 4.5 | $0.00247 | $0.00247 |
Grade B, and why
review-loop 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 11d 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, How it starts
The opening of the file, as written. The whole thing — 147 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 tools: frontmatter (root cause of the "sandbox bug")
Root cause (found 2026-09-19, fixed in v2.8.2): tools: is a list of tool names. The old values read-only and all are not tool names, so Claude Code resolved every review-loop agent to zero tools. Older Claude Code versions spawned them anyway, giving tool_uses: 0 and hallucinated output; Claude Code 2.1.278 refuses the spawn with would be spawned with zero tools — unrecognized [read-only] (or [all]). This was never a sandbox restriction on plugin agents.
Fix: read-only agents declare tools: Read, Grep, Glob, Bash (no Edit/Write; reviewer declares Read, Grep, Glob); agents that edit files (executor, code-simplifier) omit tools and inherit everything. Never put a policy word in tools: — valid values are tool names, *, or an omitted field.
Bash in report-only agents: the Bash in those frontmatters is for direct user invocation only (git diff, static analysis). The paired-session coordinator dispatches report-only roles as reviewer-role turns with that role's own permissions; see docs/protocol/reviewer-runtime.md (owner decision 2026-10-04: keep Bash, align the docs; the legacy launchers were removed in v2.13.1).
History: first seen with Executor (tools: all) in commit 8506809, then with code-simplifier (2026-04-06) and rust-reviewer (issue #3). Each time the conclusion was "plugin agent types are sandboxed", so the protocol switched to subagent_type: general-purpose with the agent body inlined in the prompt.
Rule: the paired-session coordinator runs every role as a CLI turn of its own (see Agent Invocation Pattern); no skill spawns a role through the Agent tool today. If one ever does, use subagent_type: general-purpose with the agent body inlined. Never use subagent_type: review-loop:<name>. A writable general-purpose agent cannot serve as a permission boundary for a report-only reviewer.
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.
- 11d ago Changed · +1 lines 9513898cdf0f
- 12d ago Changed · -110 lines · -1,579 tokens per session 305e27e57052
- 13d ago Changed · +8 lines · +90 tokens per session 145afdfea50f
- 14d ago Changed · +11 lines · +303 tokens per session 8610b1bbafa5
- 18d ago Changed 52078a5cd6f2
- 20d ago Changed · -1 lines · -16 tokens per session scan C → B 2bb24d56b683
- 22d ago Changed · +96 tokens per session 6e40e95b2248
- 25d ago Changed · +10 tokens per session d8e439ae0dd4
- 1mo ago Changed · +1 lines · +282 tokens per session c6059994836a
- 1mo ago First seen · 237 lines · 3,284 tokens per session scan C c0d8db6da465
review-loop CLAUDE.md is an instructions file published in the GitHub repository NYTC69/review-loop (3 stars, last pushed yesterday), licensed Apache-2.0. It adds 2,470 tokens to every session, about $0.0099 per session on Opus 5.5. A static security scan graded it B with 1 finding (strips warnings and disclaimers). 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.