security-fix — Claude Code skill: what it does, how to install
The security-fix Claude Code skill keeps vulnerability fixes in a private GitHub Security Advisory fork. It activates for security audit findings, private reports, or a pre-push rejection tied to vulnerability-related wording.
AGPL-3.0 ★ 1,120 updated 2026-11-16
npx skills add rerun-build/rerun --skill security-fix -a claude-codeWhat security-fix does
The skill routes vulnerability work through a private advisory fork instead of a public repository branch. In a separate workspace, pnpm security:start opens an advisory draft and configures its fork as the push destination. Branches sent to the public repository become visible immediately.
Run the typecheck and tests locally, then create a pull request from the fork with gh pr create. GitHub Actions do not run in advisory forks. Rerun Cloud must be deployed before the advisory is published; its Dockerfiles accept UPSTREAM_REPO and UPSTREAM_TOKEN for a private upstream.
Work in a fresh Conductor workspace or worktree on a branch that has never been pushed. The workflow also uses pnpm, gh, and GitHub Security Advisory access. The scripts/git-hooks/pre-push hook scans branch names and new commit messages for terms such as security or XSS; it is a word-based guard, and --no-verify bypasses it.
What it helps with
- patch a vulnerability without exposing a public branch
- create a private advisory fork for the fix
- run checks locally when GitHub Actions are unavailable
- prepare a fork-based pull request and deploy before disclosure
Example prompts for the agent
- Fix this vulnerability through a private advisory fork, keeping the change private until publication.
- Handle this security audit finding with the private advisory workflow.
- The pre-push hook blocked my vulnerability fix. Move the work into the private workflow.
How to install security-fix
1Skills CLI (Vercel)
npx skills add rerun-build/rerun --skill security-fix -a claude-code2Manually: put the files in place
~/.claude/skills/security-fix/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/security-fix/
1Skills CLI (Vercel)
npx skills add rerun-build/rerun --skill security-fix -a codex2Manually: put the files in place
.agents/skills/security-fix/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add rerun-build/rerun --skill security-fix -a cursor2Manually: put the files in place
.cursor/skills/security-fix/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install rerun-build/rerun .claude/skills/security-fix2Manually: put the files in place
.github/skills/security-fix/SKILL.md1Skills CLI (Vercel)
npx skills add rerun-build/rerun --skill security-fix -a gemini-cli2Platform CLI
gemini skills install https://github.com/rerun-build/rerun --path .claude/skills/security-fixManually: put the files in place
.devin/skills/security-fix/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about security-fix
How do I install security-fix in Claude Code?
Run npx skills add rerun-build/rerun --skill security-fix -a claude-code.
What does security-fix do?
It guides vulnerability fixes through a private GitHub Security Advisory fork. The workflow covers local checks, a pull request from the fork, and publication after the fix ships.
What does security-fix need to work?
Use a separate Conductor workspace or worktree, pnpm, gh, and access to GitHub Security Advisory. Deploying Rerun Cloud uses UPSTREAM_REPO and UPSTREAM_TOKEN.