Prototype · Preview: every server in the official registry, synced every 30 minutes. Reviews and accounts come in a later step.
PopNx MCPMCP servers for people who use AI · explained in plain words

Actionpin

by CloudCorpRecordsversion 0.1.8i★ –iRuns on your computer
Registry nameio.github.CloudCorpRecords/actionpin

Install

Overview

Description from the maker (original English)

GitHub Actions supply-chain hardener: audit workflows for unpinned actions and build a SHA pin plan.

PopNx MCP’s three pillars

These three are never merged into one score: opinions, measurements and popularity are different things.

What you can do with it

We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.

  • Connection: Runs on your computer
  • Transport: stdio
  • Category: Developer tools

Install

Pick the AI app you use, then copy the settings.

Before you install

  • Basic checks: 1 of 1 passed
  • Runs on your computer with your user’s permissions (files, network, apps).
  • The registry lists no key to fill in.

You need

  • npmNode.js installed (for the npx command)

Which app do you use?

Pick your app above. We remember it, so next time the button says “Install in …”.

Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json

{
  "mcpServers": {
    "actionpin": {
      "command": "npx",
      "args": [
        "-y",
        "@reneturcios/actionpin",
        "mcp"
      ]
    }
  }
}

Run this in your terminal

claude mcp add actionpin -- npx -y @reneturcios/actionpin mcp

Open in Cursor

Or copy the settings yourself

Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json

{
  "mcpServers": {
    "actionpin": {
      "command": "npx",
      "args": [
        "-y",
        "@reneturcios/actionpin",
        "mcp"
      ]
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

Open in VS Code

Or copy the settings yourself

Paste into .vscode/mcp.json in your project

{
  "servers": {
    "actionpin": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@reneturcios/actionpin",
        "mcp"
      ]
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.

Your choice stays on this device in a cookie; no account needed.

These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.

What it can do

Runs on your computer; the tool list appears in your AI app after you install it.

Basic checksi

We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 6, 2026

  • ✓ Pass

    Maker matches the repo

    The registry name was verified with the GitHub account CloudCorpRecords, which owns the repo

  • … Not checked yet

    Still maintained

    We have not read this repo from GitHub yet

  • … Not checked yet

    Has a licence

    We have not read this repo from GitHub yet

  • – Not applicable

    Endpoint responds

    This one runs on your computer; there is no internet endpoint

  • – Not applicable

    Sign-in with OAuth

    Runs on your computer and uses local settings instead

  • … Not checked yet

    No hidden instructions in tool descriptions

    Runs on your computer; we can only read tool lists from internet endpoints

Compared with others in Developer toolsi

  • More stars than 0% of the other 1384 servers in this category
  • Published to the registry more recently than 80% of the other 1384 servers in this category
  • Basic checks equal to or better than 100% of the other 1384 servers in this category

Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.

Common questions

Is Actionpin safe?

It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.

What keys or codes does it need?

The registry does not list a required key, though some servers ask you to sign in on first use.

Does it run on my computer or in the cloud?

It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).

Is it still maintained?

The GitHub repo last changed –, and the latest version (0.1.8) reached the registry 11 days ago.

Which apps can use it?

Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Node.js, Docker on your computer.

What can I use instead?

In the same category: GitHub, Desktop Commander, Cloudflare MCP

Alternatives

Servers in the same category from other makers, sorted by stars.

Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi

GitHub

by github

GitHub’s official server lets AI read and manage code, issues and pull requests in your account. Use GitHub’s hosted version or run it locally with Docker.

Computer or cloud★ 34,245MITUpdated 3 days ago
Install

Desktop Commander

by wonderwhy-er

MCP server for terminal commands, file operations, and process management

Runs on your computer★ 10,225MITUpdated 15 days ago
Install

Cloudflare MCP

by cloudflare

Cloudflare MCP servers

Runs in the cloud★ 4,446Apache-2.0Updated 14 months ago
Install

README from the maker

Open the README

Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.

actionpin

A version tag is a promise anyone can break. A commit SHA is not.

Quickstart

npx @reneturcios/actionpin audit

GitHub Actions are executable supply chain. When a workflow says uses: actions/checkout@v4, that tag can move. When a third-party action tag moves to a compromised commit, every workflow that trusts the tag runs the new code with that workflow's tokens. Pinning to a full 40-character commit SHA is the standard defense, and almost nobody does it consistently, because doing it by hand across every workflow is tedious and easy to get wrong.

actionpin audits every workflow in a repo, reports exactly which actions are pinned and which are not, and can rewrite the unpinned ones to SHAs for you, keeping the original tag as a comment so the pin stays readable.

Why this exists

GitHub Actions are the second largest confirmed open source supply-chain attack surface of the last year, behind only npm. Maintainer accounts and action tags have been hijacked before, and a mutable tag means the code your CI runs today is not guaranteed to be the code you reviewed. The fix is known and boring: pin to a SHA. actionpin makes the boring fix one command, and makes the audit callable by agents through MCP so it can run inside contribution and release workflows, not just by hand.

What it checks

For every uses: line in .github/workflows/*.yml and *.yaml:

classificationmeaning
PINNEDRef is a full 40-character commit SHA (or a full sha256 or sha512 content digest for a container image)
MUTABLERef is a version tag such as v4 or v4.2.1. The tag can move.
BRANCHRef is a branch such as main, master, or HEAD. It moves by design.
LOCALA local path such as ./local/action. Never rewritten.

Owner types: official for actions owned by actions or github, third-party for everyone else, docker for docker:// container actions, and local for local paths.

Verdict rules:

  • PINNED: every non-local action is pinned to a full SHA. Exit code 0.
  • EXPOSED: at least one third-party or container action uses a mutable tag or a branch. Exit code 1.
  • PARTIAL: only official actions are unpinned. Exit code 1.
  • Usage errors exit with code 2.

Fix mode resolves a tag or branch to its commit SHA through the GitHub API and rewrites the line to owner/repo@ # , preserving indentation and all other content. Files are edited line-wise; YAML is never parsed and re-emitted, so formatting survives. If a ref cannot be resolved, its line is left unchanged and the failure is reported. Docker image refs are reported but not rewritten; pin those by digest.

Usage

CLI
# Audit the current repo

npx @reneturcios/actionpin audit

Audit another path, machine readable

npx @reneturcios/actionpin audit ~/src/my-repo --json

Audit every repo checkout under an org root and report pin drift

(the same action used at different refs in different repos)

npx @reneturcios/actionpin org ~/src/my-org --json

Preview the pinning plan without writing

npx @reneturcios/actionpin fix --dry-run

Pin everything resolvable, in place

npx @reneturcios/actionpin fix

Example output:

actionpin audit .: EXPOSED

workflows: 2, actions: 6, pinned: 1, mutable: 4, branch: 1, local: 0 [PINNED] .github/workflows/ci.yml:12 actions/checkout@0123456789abcdef0123456789abcdef01234567 (official) [MUTABLE] .github/workflows/ci.yml:13 actions/setup-node@v4 (official) [MUTABLE] .github/workflows/publish.yml:20 docker/build-push-action@v6 (third-party) EXPOSED: at least one third-party or container action uses a mutable tag or branch ref.

Set GITHUB_TOKEN for the higher authenticated GitHub API rate limit during fix mode. Auditing needs no credentials and no network.

MCP server

Add to your MCP client config:

{

"mcpServers": { "actionpin": { "command": "npx", "args": ["-y", "@reneturcios/actionpin", "mcp"] } } }

Tools:

  • audit_workflows: audit a repo path and return every action with its classification, owner type, file, and line, plus the verdict.
  • pin_plan: return the full SHA pinning plan for a path without writing any files.
Docker
docker run -i ghcr.io/cloudcorprecords/actionpin mcp

docker run --rm -v "$PWD:/work" -w /work ghcr.io/cloudcorprecords/actionpin audit .

Limits (pinned by the hardening battery)

These behaviors are deliberate, tested, and stated plainly so a green report is never misread:

  • Scope is workflow files plus in-repo composite actions. A directory audit scans .github/workflows/*.yml and *.yaml, and also the action definitions that live inside the scanned repo: the root-level action.yml or action.yaml (the scanned repo's own published action definition, scanned whether or not a workflow references it, which is the normal shape for a repo that ships an action), every action.yml and action.yaml under .github/actions/, plus any local action directory a workflow references with a ./ uses path. Their internal steps are inventoried and classified like workflow steps, flagged as composite action refs, and drive the verdict the same way. A root action whose steps reference the repo's own action (via ./ or via its published name) is scanned exactly once; self-reference never recurses or double counts. The report lists how many composite files were scanned, and names any action files found elsewhere in the repo that were not scanned (not the root action file, outside .github/actions, and not referenced by a workflow), so coverage is never implied beyond what was audited. Passing an action.yml file path directly also parses it like any other file. Job-level uses: lines (reusable workflow calls) are in scope and are audited and pinned like step-level ones.
  • A local path with a malformed prefix is local, and the typo stays visible. A uses value such as $/.github/actions/issues-helper (a stray character where the dot of the ./ prefix belongs, shipped by real repos) names no remote repository, so it is classified LOCAL and is never resolved against the GitHub API in fix mode. The finding carries a note that the prefix is not a valid remote reference and looks like a local path typo, because the spelling is a real defect in the target repo even though it is not a pinning exposure.
  • Dynamic values are skipped, not guessed. A uses: value that contains a GitHub expression (${{ matrix.action }}, actions/checkout@${{ inputs.ref }}) is decided at run time and cannot be classified statically. actionpin skips it entirely instead of inventing an action or a verdict for it. Audit the matrix or input values themselves if a workflow builds its actions dynamically.
  • Short SHAs never count as pinned. Only a full 40-character SHA (or a full sha256 or sha512 digest for docker:// images) is PINNED. A short SHA such as abc1234 is reported in the BRANCH bucket and fails the verdict, because a short prefix is not a stable pin.
  • Container tags never count as pinned. For docker:// images only a full content digest counts. A tag spelled as 40 hex characters is still a tag and can be re-pushed, and a truncated or non-hex digest is not a digest; both are MUTABLE.
  • YAML anchors and aliases are resolved within the file. GitHub Actions has supported them since September 2025. An anchored uses: value is classified by the value itself, an alias line (uses: *name, or a step alias - *name) inherits the anchored classification and is flagged viaAlias, and anchors must be defined before use in the same file, matching GitHub. Fix mode pins the anchor definition line and never rewrites an alias line.
  • Refs containing slashes (release/v1, feature/new-thing) are reported as BRANCH bucket moving refs. Fix mode resolves them through the git-ref API with the slash preserved.
  • Block and flow style are both inventoried. Parsing is line-wise so fix mode can preserve formatting, and a quote-aware flow scanner inventories uses: entries written in YAML flow style (for example steps: [{uses: 'actions/checkout@v4'}]), including several entries sharing one line, flow collections split across lines, and quoted (JSON-shaped) keys. Commented-out uses: lines are ignored, and a uses: word inside a quoted string or a run command invents no action. Fix mode pins a flow entry without adding a version comment on the shared line, because a comment would eat the rest of that line.
  • A file that cannot be fully parsed is reported, never silent. When a workflow or composite file carries a uses: key whose value cannot be extracted, the report names the file in unparsedWorkflows, carries a note saying it could not be fully parsed, and floors the verdict at PARTIAL, so the file never silently contributes zero actions to a PINNED verdict.
  • Docker refs are reported, never rewritten. Pin container images by digest yourself; fix mode leaves docker:// lines unchanged and says so in the plan.
  • Empty estates say so. When no workflows and no action definitions are found under the audited path, the report carries a plain note that nothing was audited. The verdict is computed over zero actions and is not an all-clear for that path. A repo whose only auditable surface is its root action definition is audited on that definition and carries no empty-estate note.
  • Missing or wrong-type targets are usage errors, never a verdict. A path that does not exist exits 2 with an error on stderr and no report, matching hookcheck and injectcheck, so CI wired to a mistyped path fails instead of going green on nothing. Passing a single file audits it only when it is a .yml or .yaml file; any other regular file (a package.json, a README) is a usage error exit 2, not a workflow. A target that is neither a file nor a directory is a usage error too.
  • Non-regular files are never opened. Discovery stats every candidate before opening it and skips anything that is not a regular file, so a FIFO (named pipe), socket, or device planted at a workflow or action file path cannot hang or block the audit; the regular files around it are audited normally.
  • Piped output is delivered whole, or the run fails. Reports are written through a flush-aware emit path, so a report of any size arrives complete through a pipe, the way agents, scripts, and CI consume it. If the reader closes the pipe before the report is delivered, actionpin exits 141 (the shell-visible equivalent of SIGPIPE) instead of reporting success for output nobody received. Fixed in 0.1.2 after Round 1 real-repo validation caught silent truncation at exactly 65,536 bytes with a success exit code.

The toolkit

actionpin is part of a family of small, single-purpose developer tools by the same author, all built the same way: one bin that serves a CLI and an MCP server, deterministic checks with evidence, no invented answers.

  • greenlight: pre-flight checks for open-source contributions: is this issue actually landable?
  • shipcheck: verify a release actually shipped across git, npm, the MCP Registry, and GHCR.
  • hookcheck: audit auto-exec hooks in a repo for persistence-style risks.

Development

npm install

npm run build npm test

Tests run on fixtures with an injected fetch function; no test touches the network.

#

Only the beginning is shown; read the rest in the repo

Ask PopNx