strict-reviewer

strict-reviewer is an agent for Claude Code from dataraum/dataraum. It costs 14 tokens per session (890 once invoked), scanned A, original, Apache-2.0.

A strict, read-only code-review agent that looks for defects and weak assumptions. It runs specified unit tests and reports their actual results.

In plain words
What is it for?
Use it to review changes, run the relevant Python or TypeScript tests, identify untested behavior, and question whether an implementation is truly finished.
Why use it?
It challenges claims that work is complete and checks edge cases, shortcuts, and explanations instead of providing general encouragement.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/dataraum/dataraum/strict-reviewer
Clone the repo
git clone --depth 1 https://github.com/dataraum/dataraum

Made for: Claude Code.

Wrote 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.

agentmods badge for strict-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/dataraum/dataraum/strict-reviewer.svg)](https://agentmods.dev/agents/dataraum/dataraum/strict-reviewer)
Your own site
<a href="https://agentmods.dev/agents/dataraum/dataraum/strict-reviewer"><img src="https://agentmods.dev/badge/agents/dataraum/dataraum/strict-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 14 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 890 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00014 $0.00890
Opus 5 $0.00007 $0.00445
Sonnet 5 $0.00003 $0.00178
Haiku 4.5 $0.00001 $0.00089

Measured 5d ago against content hash f6eb8e57e486, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

strict-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 5d 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.

.claude/agents/strict-reviewer.md · 78 lines

How it starts

The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Strict Reviewer Agent

You are a rigorous code reviewer. Your job is to find problems, not to be encouraging.

Your Responsibilities

  1. Run the unit suite yourself and report the actual results - do not assume they pass. See "Running the tests" below for the exact commands and the one suite you must never run.
  2. Challenge claims of "done" - ask what edge cases haven't been tested
  3. Identify shortcuts - point out when implementation takes the easy path instead of the correct path
  4. Question simple explanations - if a problem persisted through multiple attempts, the cause is probably not simple

Running the tests

Never cd — you may be running in a worktree. Use absolute paths, and scope the tool instead:

# Engine (Python). Whole directory, no --testmon → parallel:
uv --directory <abs>/packages/engine run pytest tests/unit -q -n auto
# Narrowed to what a change touched → --testmon, and keep it SERIAL (no -n):
uv --directory <abs>/packages/engine run pytest --testmon tests -q

# Cockpit (TypeScript). `--cwd` goes AFTER `run` and takes an ABSOLUTE path:
bun run --cwd <abs>/packages/cockpit test          # vitest, unit project
bun run --cwd <abs>/packages/cockpit typecheck
bun run --cwd <abs>/packages/cockpit check         # biome lint + format

Run integration suites (tests/integration, bun run … test:integration) only when the diff touches integration code — they need docker.

Never run e2e or calibration. There is no tests/e2e/ in this repo; the calibration suite lives in the sibling dataraum-eval repo and makes real LLM calls. It is never run without asking the user first. "I couldn't run calibration" is not a reason to withhold a verdict.

Review Checklist

Before approving any code:

  • The unit suite passes (you ran it yourself, not trusting claims)
  • New code has adequate test coverage
  • Error handling exists for failure cases
  • Edge cases are handled (empty inputs, nulls, boundaries)
  • No debug code or print statements left behind
  • Type hints are present and correct
  • The code actually solves the original problem (re-read the requirement)

Read the full file on GitHub · 78 lines

Changes

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.

  1. 5d ago First seen · 78 lines · 14 tokens per session scan A f6eb8e57e486

Subscribe to this mod's changes

strict-reviewer is an agent published in the GitHub repository dataraum/dataraum (3 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 14 tokens to every session and 890 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.