Evidence record · captured 2026-10-11

GitHub Actions Hardening

A security reviewer for .github/workflows files that reasons about the Actions threat model: script injection from untrusted ${{ }} inputs, privileged triggers (pull_request_target, workflow_run, issue_comment) running fork code, mutable action references vs SHA pinning, over-scoped GITHUB_TOKEN permissions, OIDC and secret handling. Findings follow a bundled report format.

Community publisherUnverifiedSource reviewedBehavior not tested

Source review

Editorial assessment

This is a source-only assessment. It does not claim runtime or security testing.

What it does

A hardening reviewer for workflow YAML built on one insight: ${{ }} expressions are pasted into the shell before it runs. It follows a fixed seven-step audit and grades findings on five levels, down to hygiene gaps.

Best for

Auditing a repository's workflows for missing permissions blocks, mutable action tags and injection sinks
Writing a new workflow that should be locked down from the start
Explaining why a pull_request_target or workflow_run workflow is risky

Strengths in the source

Fixes the review order: map triggers and their privilege, hunt injection sinks, check privileged triggers for untrusted code, then audit permissions, action pins, and secret and output handling.
Tells the agent not to call a fork pull_request dangerous just for running untrusted code, since it gets no secrets and a read-only token.
Every CRITICAL or HIGH finding must come with a corrected YAML snippet, and changes are presented for review, never applied automatically.
Five reference files cover the trust matrix per trigger, safe env-variable patterns, least-privilege recipes with OIDC, pinning with Dependabot, and the report template.

Limitations

Static YAML review only; it does not inspect repository settings such as the default token permission or branch protections, which change the real risk.
Takes a stricter line than exploit-focused reviewers: a missing permissions block is MEDIUM and tag-pinned first-party actions are LOW, which can produce findings no outsider can exploit.
Advice reflects GitHub Actions behavior at the time of writing and may need updating as GitHub changes defaults.

When not to use it

When you only want findings an outside attacker can actually exploit; an exploit-gated review fits better
CI systems other than GitHub Actions

Reviewed by Retenta editorial on 2026-10-11; scope: source only.

Files included in this review

skills/github-actions-hardening/SKILL.md

Evidence and boundaries

Execution · not tested

Retenta has not executed this skill.

Security · not tested

The skill folder bundles no executable scripts. Retenta has not run a security scan.

Pinned publisher source: skills/github-actions-hardening/SKILL.md
Revision: 82701c24b99488536ca399ff4789a458b7a05db7
License evidence: MITRetenta preserves the license text captured with this immutable upstream snapshot; no independent legal or security review is claimed.
Last source change: 2026-06-16
Independent verification: Published by GitHub awesome-copilot (contributed by AmirhosseinHonardoust), which does not own the product this skill covers. Not independently verified.
Upstream repository: Open publisher repository

Compatibility

GitHub Copilot (VS Code / Copilot CLI)

documented

Named in the upstream skill or repository documentation; not tested by Retenta.

Requirements and access

None required (static review)

Referenced tools

No named tool is required by the reviewed SKILL.md.

Install from the source

Clones the current default branch, which may be newer than the reviewed revision 82701c2. The destination shown is Claude Code's personal skills folder; other agents read the same folder from their own skills directory.

View the SKILL.md at its source Verify the publisher's install documentation

More Security skills →

Install the full skill folder

git clone --depth 1 https://github.com/github/awesome-copilot.git
cp -R awesome-copilot/skills/github-actions-hardening ~/.claude/skills/github-actions-hardening

Thyme RAG Capsules

Make this skill’s state durable, governed, and portable.

A Retenta capsule lets agents recall prior work, append provenance-stamped progress, and carry the memory to the next run.

Create durable agent memory Understand the evidence labels