- Home
- Skills
- DevOps & Cloud
- github actions workflow auditor
More screenshots
Works with the AI Agents you already use
github actions workflow auditor
Audit GitHub Actions YAML for write-all permissions, pull_request_target, unpinned actions, script injection from github.event.* and secrets on untrusted checkout.
$7
github actions workflow auditor
Example session with this skill installed
Audit the workflow at .github/workflows/ci.yml and tell me if it is safe to run on pull requests from forks.
- Read your context and instructions
- Compiled the github actions workflow
I have analyzed .github/workflows/ci.yml. The audit identified 2 HIGH severity risks related to trust boundaries.
GitHub Actions Security Audit: ci.yml
| Severity | Category | Finding | Recommendation |
|---|---|---|---|
| HIGH | Injection | github.event.pull_request.title used in run | Intermediate to an env variable before use in shell. |
| HIGH | Trigger | pull_request_target with actions/checkout | Use pull_request or remove write permissions. |
| MEDIUM | Pinning | actions/setup-node@v3 is not pinned to a SHA | Use actions/setup-node@8f152de45cc393bb60ce5cd0fb1ad5441 |
| LOW | Permissions | Top-level permissions block missing | Add permissions: contents: read to the workflow. |
Summary of Risks
The workflow uses pull_request_target while checking out code from the incoming branch. An attacker could submit a PR with a modified package.json that executes malicious code with your repository's write tokens. Additionally, the job name or PR title is passed directly to the shell, allowing for command injection if a PR is named carefully.
Next steps
- Update
ci.ymlto use explicitpermissions:at the job level. - Replace mutable tags (like
@v3) with full 40-character commit SHAs. - If you need to scan for leaked credentials, I can run a secrets pattern scanner.
Connects securely to your tools. The creator never sees your data.
About this skill
Find dangerous trust boundaries in your workflows before a malicious PR does.
GitHub Actions is a common supply-chain entry point. This skill heuristically audits workflow YAML for permissions: write-all, pull_request_target, actions pinned to @main/@master, script injection from github.event.* into run:, and secrets combined with untrusted checkout.
What it does
- Scans one workflow or a whole
.github/workflowsdirectory - Flags mutable action refs and recommends full commit SHAs
- Suggests least-privilege top-level
permissions: - Markdown + CSV; exit 0 when there are no HIGH findings
Who it's for
Platform / DevOps engineers maintaining CI for public or open-contribution repos.
Limits
Heuristic text scan (no full YAML schema). Still review environment protection and OIDC. Not legal advice.
What's included
SKILL.md: workflow, commands, how to explain findingsscripts/: the checker (stdlib Python) plus sharedseccommon.pyreferences/sources.md: specs and guidance citedexamples/: good and bad samples used in the demotests/: unit tests (Python 3.8 / 3.9 / 3.13)
Standard-library Python only. No API keys, no network calls, no installs.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for DevOps & Cloud.
- Offline / local files only — does not fetch live URLs or call cloud APIs.
- Heuristic checks; not a full substitute for a professional audit or pentest.
- Not legal advice and not a certification of compliance.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 20 days ago
Permissions
File Scopes
Fully offline. No subprocess, no network modules used for outbound calls, no credential access.