patch

Generate a minimal in-repo patch for a security finding and output a unified diff.

Updated May 20, 2026
One-click install
npx skills add https://github.com/lukehinds/scrutineer-attest-test --skill patch-lukehinds
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: patch
Source: https://github.com/lukehinds/scrutineer-attest-test/tree/main/skills/patch
Command: npx skills add https://github.com/lukehinds/scrutineer-attest-test --skill patch-lukehinds

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Propose a minimal, review-ready patch for a security finding by editing the codebase in place and returning a unified diff plus a concise rationale, enabling analysts to review and apply changes without automated pushes.

Core Features & Use Cases

  • Minimal, surgical patches targeted at the reported finding that fix the vulnerability without broad refactoring.
  • In-place edits against the HEAD repository (./src) guided by context.json to locate the sink and vulnerability trace.
  • Outputs a patch diff (patch.diff) and a briefing note for the finding to accompany reviewer review and PR preparation.

Quick Start

Provide the context.json with a finding_id so the skill can fetch the finding and generate a patch diff with rationale.

Frequently Asked Questions about patch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a minimal patch for a security finding without broad refactoring?

A unified diff patch fixes security findings by editing the codebase in place against HEAD, guided by context.json to locate the sink and produce a minimal patch without broad refactoring.

What is a unified diff and how does it help with code review for security fixes?

A unified diff is a text format showing changes made to files, enabling analysts to review surgical security patches and rationale before applying changes or preparing a pull request.

Can I automatically push security patches directly to my repository?

No, automated pushes are not supported. The Skill outputs a patch diff and a briefing note for manual review and PR preparation, ensuring analysts validate changes before applying them.

How do I trace a vulnerable data flow to create a targeted security fix?

Provide context.json with a finding_id to fetch the finding, locate the sink in ./src, trace the vulnerable data flow, and produce a targeted patch with a concise rationale.

Does this patch generation approach work with any Git repository HEAD?

Yes, it operates on the repository HEAD and uses ./src with context.json to locate the finding's location, making it compatible with any Git repository structure.