fix-ci

An agent skill by cursor, from cursor/plugins. Tags: ci-cd, debugging, github-actions, troubleshooting, validation.

What it does

Find failing PR checks, inspect logs or external check links, and apply focused fixes

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add cursor/plugins --skill fix-ci

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/cursor/plugins
cp -r plugins/cursor-team-kit/skills/fix-ci ~/.claude/skills/fix-ci

Safety box score

Grade C, moderate risk: 26 of 100 risk points, rated 2026-10-16. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill is a GitHub- and CI-focused fixer: it reads PR checks and logs, inspects external check links, edits the repo, and pushes iterative fixes back to the branch. The main risks are remote network access to GitHub/other check services and autonomous code pushes until checks turn green.

Category levels, from 0 (none) to 3 (high)
Category	Level	Why
Code & shell execution	1 (low)	It drives shell/CLI activity via `gh pr checks --json name,bucket,state,workflow,link` and log inspection, but there are no bundled scripts or code execution hooks.
Network access	2 (moderate)	It depends on GitHub APIs, Actions logs, and possibly external check links ('inspect logs or external check links', 'use the check link').
Filesystem reach	1 (low)	It applies changes in the repository working tree ('Apply the smallest safe fix'), but does not describe reads or writes outside the project.
Secrets & credentials	1 (low)	It assumes authenticated GitHub CLI access and may expose CI log contents, but it does not ask for passwords, tokens, or key files directly.
Privilege & persistence	0 (none)	No sudo, global installs, service registration, or OS/shell configuration changes are mentioned.
Instruction hijack surface	0 (none)	There is no text telling the agent to ignore safeguards, hide actions, or follow remote instructions.
Irreversible actions	2 (moderate)	The workflow explicitly says to 'Push, re-check the PR check set, and repeat until green,' which authorizes repeated remote writes without a human checkpoint each iteration.
Transparency	0 (none)	The skill file is plain markdown only; there are no binaries, encoded payloads, or hidden installers.

Before installing

Check that the GitHub token/`gh` account it will use is limited to the intended repo or PRs.
Confirm you are comfortable with the agent pushing commits to the branch without a fresh approval each iteration.
Review whether your CI logs or external check links may reveal sensitive build output.

Rated 2026-10-16 by an automated review of 1 file from the skill's repository.

Source

Repository: cursor/plugins (all skills from this repository)
Path: cursor-team-kit/skills/fix-ci/SKILL.md
Branch: main
Collection: cursor
Updated: 2026-09-23

debug-voice — Use when the user runs /debug-voice, says voice mode has flaws, asks to see or capture what happened in a Grok realtime voice session.
loop-on-ci — Monitor PR checks and fix failures until green. Uses gh pr checks as the source of truth for PR-attached checks.
principle-fix-root-causes — Apply when debugging.
reproduce-and-fix-issues — Reproduce triaged Slack bugs through a configured app-control adapter, verify existing fixes.
claude-skills-troubleshooting — Diagnose and resolve Claude Code plugin and skill issues.
debugging-network-issues — Investigates network, streaming, and protocol failures with falsification-first layered experiments.