do-code-review
Review the branch you are on since a fixed point, on six Axes (correctness, spec fidelity, repo standards, principles, blast radius, security), every Finding…
Skill by Thalisu
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/thalisu-do-code-review
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: security
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Shell scripts
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/do-code-review
What's inside
- Skill name
do-code-review- Description
- Review the branch you are on since a fixed point, on six Axes (correctness, spec fidelity, repo standards, principles, blast radius, security), every Finding proven to a Rung and written by Bucket into one Review file beside the branch. Use when the user asks to review this branch, review the diff, review since a ref, check the branch before a push, or says 'revisa esse diff', 'revisa essa branch', 'review since main'; also the review step of do. Do not use for a pull request the user wants reviewed and posted on GitHub: that is the bundled /code-review, which this skill leaves untouched.
- License
- MIT
- Instructions
- 4 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (147)
- AGENT.md
- SKILL.md
- do-code-review-fixer.md · agents/
- do-code-review-gate-fixer.md · agents/
- do-code-review-security-reviewer.md · agents/
- do-code-review-spec-reviewer.md · agents/
- do-code-review-technical-reviewer.md · agents/
- openai.yaml · agents/
- README.md · evals/
- case.yaml · evals/empty-diff/
- nothing-written.md · evals/empty-diff/graders/
- one-line-refusal.md · evals/empty-diff/graders/
- prompt.md · evals/empty-diff/
- case.yaml · evals/fix-after-review-forks-no-fixer/
- finding-named-ahead-of-the-red-gate.md · evals/fix-after-review-forks-no-fixer/graders/
- finding-named-in-the-landing.md · evals/fix-after-review-forks-no-fixer/graders/
- no-fixer-forked.md · evals/fix-after-review-forks-no-fixer/graders/
- not-landed-and-nothing-edited.md · evals/fix-after-review-forks-no-fixer/graders/
- prompt.md · evals/fix-after-review-forks-no-fixer/
- case.yaml · evals/fix-already-fixed/
- earlier-lines-kept.md · evals/fix-already-fixed/graders/
- green-not-blocked.md · evals/fix-already-fixed/graders/
- no-fixer-forked.md · evals/fix-already-fixed/graders/
- settled-verified.md · evals/fix-already-fixed/graders/
- prompt.md · evals/fix-already-fixed/
- case.yaml · evals/fix-dirty-tree/
- nothing-written.md · evals/fix-dirty-tree/graders/
- one-line-refusal.md · evals/fix-dirty-tree/graders/
- prompt.md · evals/fix-dirty-tree/
- case.yaml · evals/fix-duplication-scan-stays-dirty/
- still-dirty-lands-nothing.md · evals/fix-duplication-scan-stays-dirty/graders/
- prompt.md · evals/fix-duplication-scan-stays-dirty/
- case.yaml · evals/fix-duplication-scan/
- duplicate-promoted-and-landed.md · evals/fix-duplication-scan/graders/
- scan-block-holds-diff-rows-alone.md · evals/fix-duplication-scan/graders/
- scan-line-recorded.md · evals/fix-duplication-scan/graders/
- scan-once-per-check-pass.md · evals/fix-duplication-scan/graders/
- scan-rows-reach-gate-fixer.md · evals/fix-duplication-scan/graders/
- prompt.md · evals/fix-duplication-scan/
- case.yaml · evals/fix-gate-fixer-no-return/
- gate-fixer-forked-once-then-no-return.md · evals/fix-gate-fixer-no-return/graders/
- landing-names-the-non-return-beside-the-finding.md · evals/fix-gate-fixer-no-return/graders/
- prompt.md · evals/fix-gate-fixer-no-return/
- case.yaml · evals/fix-gate-red/
- gate-fixer-forked-by-name.md · evals/fix-gate-red/graders/
- gate-fixer-holds-no-review.md · evals/fix-gate-red/graders/
- landed-after-gate-fixer.md · evals/fix-gate-red/graders/
- scan-skipped-no-scan-named.md · evals/fix-gate-red/graders/
- prompt.md · evals/fix-gate-red/
- case.yaml · evals/fix-no-check/
- fixer-forked.md · evals/fix-no-check/graders/
- never-verified-without-a-check.md · evals/fix-no-check/graders/
- prompt.md · evals/fix-no-check/
- case.yaml · evals/fix-parallel-waves/
- one-commit-per-finding.md · evals/fix-parallel-waves/graders/
- re-routed-fixer-forked-twice.md · evals/fix-parallel-waves/graders/
- re-routed-once-then-fixed.md · evals/fix-parallel-waves/graders/
- wave-obeyed.md · evals/fix-parallel-waves/graders/
- prompt.md · evals/fix-parallel-waves/
- case.yaml · evals/fix-stale/
- …and 87 more
SKILL.md (first lines)
Run the do-code-review contract for these arguments, in the mode they name: a `fix` call reviews nothing, fixes the Review it names and ends with the push command; `--no-fix` writes the Review and stops, ending with its text and its location; every other call reviews the diff, then fixes and lands, and ends with the Review's text, its location and the push command, or with its outcome alone and the push command when the Gate or a `Caller: do` line was handed, as only `do` does. $ARGUMENTS
From Thalisu/SKILLS, MIT.
More from Thalisu/SKILLS
- 0.592score
how
Thalisu
Explains how a subsystem works: explores the codebase and produces a senior-engineer walkthrough of its architecture, runtime flow, key concepts, where things…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
architect
Thalisu
Design the shape before code: sketch the caller's usage first, then the types, function signatures and module boundaries with `not implemented` bodies…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
improve-prompt
Thalisu
Rewrite a prompt for one Claude model from Anthropic's live prompting docs for that model: every change cites the section of the docs that asks for it, what…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
journey
Thalisu
Walk every path of a spec from the actor's seat: how they arrive, what they see, what they do, what the system answers, what happens when it fails.
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
debt
Thalisu
Gather a feature's Pending debt (its Debt files and the Review Findings still open) and judge it into a Debt review beside the Spec
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
do
Thalisu
Match a request to one Playbook and run its steps: a Ticket's path or issue reference builds that Ticket as the last step of the chain, a request in words…
SkillAFreshness: A★ 1Claude Code +9 - 0.588score
tickets
Thalisu
Cut a spec, and the journey its verdict points at, into tracer-bullet tickets: one demoable vertical slice per ticket, or a Logic ticket and a Front-end…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
discuss
Thalisu
Interview the user about a plan before code is written: one question at a time, each with a recommended answer and the failure it hunts, answered from the…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
discover
Thalisu
Batch "does this already exist in the repo / where is it" lookup, answered in one terse line per item by the discover agent (FOUND / DUPLICATE / PARTIAL /…
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
setup-pre-commit
Thalisu
Set up a Husky pre-commit hook in the current project: finds where the hook belongs, keeps the formatter the project already uses instead of forcing Prettier…
SkillAFreshness: A★ 1Claude Code +9 - 0.583score
sketch
Thalisu
Settle the shape a piece of work has to hold before any logic and file it: the caller's usage, the types, the signatures and the module boundaries with…
SkillAFreshness: A★ 1Claude Code +9 - 0.582score
teach
Thalisu
Explain a change or subsystem plainly so the user genuinely understands it instead of getting a summary.
SkillAFreshness: A★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 20, 2026 (9 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/Thalisu/SKILLS
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from Thalisu
- 0.592score
journey
Thalisu
Walk every path of a spec from the actor's seat: how they arrive, what they see, what they do, what the system answers, what happens when it fails.
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
architect
Thalisu
Design the shape before code: sketch the caller's usage first, then the types, function signatures and module boundaries with `not implemented` bodies…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
how
Thalisu
Explains how a subsystem works: explores the codebase and produces a senior-engineer walkthrough of its architecture, runtime flow, key concepts, where things…
SkillAFreshness: A★ 1Claude Code +9 - 0.592score
improve-prompt
Thalisu
Rewrite a prompt for one Claude model from Anthropic's live prompting docs for that model: every change cites the section of the docs that asks for it, what…
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
do
Thalisu
Match a request to one Playbook and run its steps: a Ticket's path or issue reference builds that Ticket as the last step of the chain, a request in words…
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
debt
Thalisu
Gather a feature's Pending debt (its Debt files and the Review Findings still open) and judge it into a Debt review beside the Spec
SkillAFreshness: A★ 1Claude Code +9
Cite this page
do-code-review skill. Agent Mart. https://agentmart.dev/skills/thalisu-do-code-review (last updated 2026-10-24)