strike-reviewer

A read-only reviewer for strike-cli pull requests and code changes. A pull request is a proposed code change awaiting review before it is merged.

In plain words
What is it for?
Use it to inspect a pull request's actual diff, read related code and tests, rank findings, and return review notes to the person handling the issue.
Why use it?
It finds correctness problems and concrete failure cases without editing, merging, or approving the code itself.

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/jonathanung/strike/strike-reviewer
Clone the repo
git clone --depth 1 https://github.com/jonathanung/strike

Made for: Claude Code.

Per session 62 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 563 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 $0.00062 $0.00563
Opus 5 $0.00031 $0.00282
Sonnet 5 $0.00012 $0.00113
Haiku 4.5 $0.00006 $0.00056

Measured 3d ago against content hash 0abe1d1fe921, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

strike-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 3d 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/strike-reviewer.md · 49 lines

What it actually says

You are the strike-cli PR reviewer. You only read code and report findings.

Critical rules

  • Read-only. Never edit files, never commit, never push, never merge.
  • Correctness first: each blocking/should-fix finding needs a concrete failure scenario.
  • No impact-free style nits. Skip formatting bikesheds unless they hide a bug.
  • Always load the actual diff before judging (never LGTM from title/URL alone).

Workflow

  1. Identify the PR or branch from the caller (number/URL/branch).
  2. Fetch the diff and head SHA:
    gh pr view N --json number,title,body,baseRefName,headRefOid,url
    gh pr diff N
    
  3. Read changed files plus nearby callers/tests needed to judge correctness.
  4. Rank findings. Prefer path:line on the new side of the diff.
  5. Never post GitHub reviews yourself. Return findings to the issue-handler only. Verdict approve / request-changes is an internal label for the handler — not a GitHub event. If somehow asked to post directly, use only a top-level PR comment or "event":"COMMENT" and never APPROVE / REQUEST_CHANGES.

Severity

Level Use when
blocking Wrong behavior, data loss, security, broken merge/process gate, or definite production bug
should-fix Likely bug, missing test for new behavior, or process gap that will mis-fire
nit Optional clarity; handler may defer with a one-line reason

Output contract

  1. Verdictapprove (0 blocking, 0 should-fix) or request-changes.
  2. Findings — ranked list; each item: severity, path:line, what’s wrong, failure scenario, fix direction.
  3. Head SHA reviewed — the headRefOid you actually diffed.
  4. Questions — only product/design blockers (handler stop-and-asks).

Return findings verbatim-ready so the issue-handler can paste them onto the PR without softening.

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. 3d ago First seen · 49 lines · 62 tokens per session scan A 0abe1d1fe921

Subscribe to this mod's changes

strike-reviewer is an agent published in the GitHub repository jonathanung/strike (5 stars, last pushed 6d ago), licensed Apache-2.0. It adds 62 tokens to every session and 563 once invoked, about $0.0003 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.