Action Audit logo

Action Audit

Organization
bitwarden
action-audit

Audit GitHub Actions action usage across an org. Searches for a specific action (incident mode) or sweeps all workflow files for non-compliant action references (audit mode). Produces a read-only report of findings with compliance status and resolved SHAs. Does not modify any files. <example> User: We need to check if any repos are using tj-actions/changed-files Action: Trigger action-audit in incident mode for that action </example> <example> User: Can you find all unpinned actions across the org? Action: Trigger action-audit in audit mode </example>

Overview

Publisherbitwarden
Repositoryai-plugins
Skill nameaction-audit
Stars
149
Forks
19
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by bitwarden on GitHub. Read the source before you install it.

Installation

Install the Action Audit AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/bitwarden/ai-plugins.git /tmp/ai-plugins
mkdir -p .claude/skills
cp -r /tmp/ai-plugins/plugins/bitwarden-devops-engineer/skills/action-audit .claude/skills/action-audit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Action Audit in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Action Audit on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Action Audit is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Rules

  • This skill is strictly read-only. Do not modify, create, or delete any files.
  • No mutating API calls. gh api GET requests are allowed freely. Do not use -X POST, -X PUT, -X PATCH, or -X DELETE.
  • Flag uncertainty. If a finding is ambiguous, note it in the report rather than guessing.

Pin Compliance Rules

Before classifying any action reference, read ${CLAUDE_PLUGIN_ROOT}/skills/bitwarden-workflow-linter-rules/SKILL.md and apply the step_pinned rule as the compliance definition for all steps below. That skill is the single source of truth for what is and is not compliant.

Modes

  • incident (default): Targeted search for a specific action — used when an action is compromised or deprecated.
  • audit: Sweep all workflow files org-wide for any non-compliant action references.

Step 1: Parse Context

Determine the mode from the user's request:

  • If the user names a specific action (e.g., tj-actions/changed-files), use incident mode.
  • If the user asks for a general sweep of unpinned actions, use audit mode.
  • If a replacement action is mentioned, note it for the remediation step (handled separately by the action-remediate skill).

Step 2: Search Org-Wide

Incident mode — search for the specific action:

bash
gh search code "uses: <action-name>" --owner <org> --path .github/workflows/ --limit 100

Also search without the uses: prefix to catch indirect references:

bash
gh search code "<action-name>" --owner <org> --path .github/workflows/ --limit 100

Audit mode — find all workflow files and extract uses: references:

bash
gh search code "uses:" --owner <org> --path .github/workflows/ --limit 100

Then apply the step_pinned compliance filter from ${CLAUDE_PLUGIN_ROOT}/skills/bitwarden-workflow-linter-rules/SKILL.md to each reference.

Note: GitHub code search indexes can lag by minutes to hours after a recent push. Results may not reflect the very latest commits. Flag this caveat in the output.

Step 3: Parse and Display Results

For each uses: reference (excluding local ./ paths), determine:

  1. Repo and file path
  2. Current uses: value (full line)
  3. Action type:
    • internal — starts with bitwarden/
    • third-party — all others (excluding local)
  4. Pin status:
    • hash — pinned to a full 40-char SHA
    • tag — pinned to a version tag (e.g., @v3, @v1.2.3)
    • branch — pointing to a named branch (e.g., @main, @master)
    • none — no ref at all
  5. Compliant: Apply the step_pinned rule from ${CLAUDE_PLUGIN_ROOT}/skills/bitwarden-workflow-linter-rules/SKILL.md — ✅ if compliant, ❌ otherwise.

Display a table:

RepoFileCurrent ReferenceTypePin StatusCompliant
..................

In incident mode, include all rows. In audit mode, omit compliant (✅) rows.

If there are no non-compliant findings, inform the user and stop.

Step 4: Resolve Remediation Targets

Apply the correct fix approach based on action type and mode. Do not treat all non-compliant references the same way.

Incident mode — replacement action provided:

If the user mentioned a replacement action in Step 1, do not resolve a SHA for the compromised action. Instead, resolve the SHA for the replacement action:

bash
gh api repos/<owner>/<repo>/commits/<ref> --jq '.sha'

Present the resolved replacement SHA and a verification link (https://github.com/<owner>/<repo>/commit/<sha>) to the user. Ask for confirmation before finalizing.

Internal actions (bitwarden/):

  • The expected fix is to change the ref to @main. No SHA resolution needed.
  • If the action is currently on a SHA, do not automatically treat this as non-compliant — a SHA pin is more restrictive than @main and may be intentional (e.g., frozen during a security incident or pinned for reproducibility). Inform the user and ask whether to change it to @main before including it in the remediation list.

Third-party actions:

  • Resolve the current SHA for each unique non-compliant action:
bash
gh api repos/<owner>/<repo>/commits/<ref> --jq '.sha'

Where <owner>/<repo> is the action's repo and <ref> is the target tag or main.

Present to the user:

  • Resolved SHA
  • Verification link: https://github.com/<owner>/<repo>/commit/<sha>

Ask: "Does this SHA look correct? Type yes to confirm, or provide a different SHA."

Wait for confirmation before finalizing the report.

In audit mode, group unique third-party actions and resolve each once rather than per-occurrence.

Step 5: Summary Report

Output a final summary:

RepoFileCurrent ReferenceTypeCompliantRemediation
..................

The Remediation column should contain:

  • For internal actions: change ref to @main
  • For third-party actions: the resolved 40-char SHA + inline comment to add (e.g., @abc123...def456 # v4.1.1)

Inform the user that they can use the action-remediate skill to apply fixes based on these findings.

Frequently asked questions

What does the Action Audit AI skill do?

Audit GitHub Actions action usage across an org. Searches for a specific action (incident mode) or sweeps all workflow files for non-compliant action references (audit mode). Produces a read-only report of findings with compliance status and resolved SHAs. Does not modify any files. <example> User: We need to check if any repos are using tj-actions/changed-files Action: Trigger action-audit in incident mode for that action </example> <example> User: Can you find all unpinned actions across the org? Action: Trigger action-audit in audit mode </example>

Why use Action Audit on TypingMind?

Because you install it once and use it with any model. Action Audit is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Action Audit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/bitwarden/ai-plugins/tree/main/plugins/bitwarden-devops-engineer/skills/action-audit. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Action Audit?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Action Audit?

As many as you like. As long as a model supports skills, you can use Action Audit with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Action Audit AI skill free?

It is published on GitHub by bitwarden. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇