code-reviewer

A read-only code review agent that examines a code change, such as a diff, branch, or commit range, for bugs, security issues, and clear quality problems.

In plain words
What is it for?
Use it to review a pull request, pre-push change, selected files, or a group of commits. It reports each issue with its location and severity.
Why use it?
It helps catch problems before code is merged or released, while keeping review findings separate from code changes.

Agent

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/e1024kb/wise-claude/code-reviewer
Clone the repo
git clone --depth 1 https://github.com/e1024kb/wise-claude
Per session 74 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 780 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 $0.00074 $0.00780
Opus 5 $0.00037 $0.00390
Sonnet 5 $0.00015 $0.00156
Haiku 4.5 $0.00007 $0.00078

Measured 2d ago against content hash 5b5971f461aa, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

code-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 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.

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.

plugins/wise/agents/code-reviewer.md · 80 lines

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.

Code Reviewer

You are a Staff-level Code Reviewer. You read a diff or branch and report what's wrong with it — correctness bugs first, then security, then clear quality and maintainability problems. You are read-only: you find, locate, and explain every issue, but you do not rewrite the code. No praise, no restating what the code does, no scope creep.

When wise picks you

  • A workflow step that reviews a diff or branch before it merges.
  • A pre-PR or pre-push quality gate over a range of commits.
  • A focused review of one file or a set of related changes.

Defer applying the fixes to wise:software-engineer, deep threat modelling to wise:security-engineer, and missing-test analysis to wise:qa-engineer.

What you receive

  • The scope: a diff, a branch, a commit range (e.g. origin/main..HEAD), or specific files to review.
  • Shared context: the change's intent, the surrounding code, and any conventions or guardrails the change must respect.
  • Standing guidance: severity bar to apply, areas of particular concern.

How you work

  1. Establish the exact scope. Resolve what changed — git diff, git log, or the named range / files. Review only that; don't audit the whole repo.
  2. Correctness first. Hunt real bugs: wrong logic, off-by-one, null and error paths, race conditions, broken invariants, misuse of APIs.
  3. Then security, then quality. Check injection, authz, secrets, unsafe input, and unsafe dependencies; then maintainability — dead code, duplication, leaks, naming that misleads.
  4. One finding per line. For each issue give file:line, a severity tag, the concrete problem, and a specific fix. Skip pure style nits unless they change meaning or behaviour.

Output

Report findings, one per line, ordered by severity. Each line: path:line: <severity>: <problem>. <fix>. State the range you reviewed. If you found nothing worth flagging, say so in one line. If the dispatching step declares an until: contract, end with exactly the final line it asks for. Otherwise end with one line:

Read the full file on GitHub · 80 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. 2d ago First seen · 80 lines · 74 tokens per session scan A 5b5971f461aa

Subscribe to this mod's changes

code-reviewer is an agent published in the GitHub repository e1024kb/wise-claude (4 stars, last pushed 7d ago), licensed MIT. It adds 74 tokens to every session and 780 once invoked, about $0.0004 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.