LuluSkill
EN
Language

Language and theme are saved in this browser only.

Start with a goal
DevelopmentPinned at c68487d

Evidence-based code review

Review code for actionable defects by tracing agreements across boundaries, constructing a feasible failure and checking the strongest counterargument. Correctness is the default; performance and security are optional review dimensions.

phuryn on GitHubUpdated Oct 13, 2026MIT
Prepare my task

Advertising pauses before you enter or manage personal content.

How to use it

Advanced. Specify the repository or diff, fixed baseline and review dimensions. Ask for findings with expected behaviour, feasible trigger, source locations, consequence and checked counterargument. The original can delegate read-only reviews and run probes, so approve repository access and command execution. No edit is requested by default. Static findings are not confirmed exploits or measured performance regressions; specialized commands need their original sibling skills.

The instructions themselves are not copied here. They stay in the author’s repository at the version below, so open them there and follow the author’s own steps.

Open the instructions at this version

What it asks your AI tool to do

  • Read files in your project or the files you point it to
  • Run commands on your computer

This is what the instructions ask for, not something LuluSkill does. In a chat window nothing runs; in an AI tool that can work with your files, what actually happens depends on that tool and its permission settings.

Source, licence and checks

Written for (as the author states)Claude Code, Claude Cowork
Checked as a prompt forCodex, Claude. The instructions can be given to these tools as a prompt; LuluSkill did not run them.
Licence and authorMIT licence, by phuryn on GitHub. If you copy or share the instructions, keep the author’s copyright notice and the licence text with them.
Pinned versionThis page describes commit c68487d721e4 of pm-ai-shipping/skills/code-review/SKILL.md. A newer version in the author’s repository is not covered until it is reviewed again.
Automated scanPassed (lowest risk), on Oct 13, 2026: 5 files checked, 0 findings. The scan looks for patterns such as shell commands, web requests, credential access and instruction overrides. It does not run the Skill and is not a full security audit.
ReviewRecorded on Oct 13, 2026 by LuluSkill’s AI project supervisor on the owner’s behalf. It checked the source, the licence file, the scan result and this description. It is not a human review and not a legal opinion.

Advertising pauses before you enter or manage personal content.

Community notes

0

Sign in to add a note.

No notes yet. Be the first to add one.

Report a problem or request removal

Rights holders and users can report licence, attribution, security or accuracy problems with this page. We respond within 48 hours and record every request in our review log.

Opens your email app. You can also write to contact@luluskill.com and include this page’s address.