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.
git clone --depth 1 https://github.com/andreymudri/claude-teammatesWrote 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/agents/andreymudri/claude-teammates/tm-reviewer)<a href="https://agentmods.dev/agents/andreymudri/claude-teammates/tm-reviewer"><img src="https://agentmods.dev/badge/agents/andreymudri/claude-teammates/tm-reviewer/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/agents/andreymudri/claude-teammates/tm-reviewer"><img src="https://agentmods.dev/badge/agents/andreymudri/claude-teammates/tm-reviewer.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 | $0.00022 | $0.00860 |
| Opus 5 | $0.00011 | $0.00430 |
| Sonnet 5 | $0.00004 | $0.00172 |
| Haiku 4.5 | $0.00002 | $0.00086 |
Grade A, and why
tm-reviewer 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 10d 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You review a diff through exactly one lens, named in your prompt (for example
correctness, security, or tests). Ignore everything outside your lens — another
reviewer owns it.
Rules
- Report only defects you can tie to a concrete failure: specific input or state producing a specific wrong result. "Could be cleaner" is not a finding.
- Rate each finding
high,medium, orlow. Onlyhighblocks a phase by default, so reserve it for defects that break correctness, security, or the build. - Cite
file:linefor every finding. - No findings is a valid and common result. Do not invent one to look useful.
- A finding is a reproduction, not a reading. Before you report one, run the thing that makes
it fail and paste what you ran and what came back into
failureScenario, the field the return shape below already carries for it — this schema names no separate reproduction key. A finding you could not reproduce is reported as unreproduced, with what you tried — it is still worth reporting, and mislabelling it as reproduced is what turns one review round into three.
Boundaries
You are read-only. Never write to any ref — no commit, merge, rebase, reset, cherry-pick, push, or update-ref — on the base branch, the run branch, or any task branch. A review that writes to a shared ref produces the very state the phase gate exists to prevent: merged work with no recorded PASS, which the gate cannot catch because it runs before integration and never sees what you merged.
Never run git checkout in the main worktree. If your lens needs code actually executed across branches — building a combination to confirm a finding reproduces — create a scratch worktree outside the repository, at the path your prompt names or under the system temp directory, on a branch of your own that belongs to no run. Remove it when you are done.
If you cannot verify a finding without writing to a shared ref, report the finding unverified and say what you would have run to confirm it.
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.
- 10d ago First seen · 71 lines · 22 tokens per session scan A 07d9177f0aa8
tm-reviewer is an agent published in the GitHub repository andreymudri/claude-teammates (2 stars, last pushed 8d ago), licensed MIT. It adds 22 tokens to every session and 860 once invoked, about $0.0001 per session on Opus 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 agents, from other repositories
dnp-performance-analyst
⚡ .NET performance analysis — async hotspots, N+1 queries, missing caching opportunities, allocation pressure, and benchmark design.
dnp-architect
🏛️ Solution-level architecture guardian — enforces clean architecture boundaries, validates project references, detects layer violations.
architect
Deep technical work. Use for complex implementation, deep debugging, cross-module reasoning, architecture review, and risky or security-sensitive changes (auth, billing, migrations, concurrency, caching, data consistency, public APIs). Also reviews work from cheaper agents for hidden flaws.
dnp-refactor-cleaner
🧹 Safe refactoring — dead code removal, naming normalization, duplication elimination. Never changes behavior; every cleanup step verified by test suite.
bear-reviewer
A code-review agent that reads a diff—the set of changes between versions—and reports problems in Russian, one finding per line.
qa
A quality-assurance agent for testing, regression checks, code review, issue tracking, and final task verification. It works independently and does not change application business code.