More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Publish Pre-flight: make it public without leaking your day

    1

    Audit Git history and files for secrets, employer names, and private data before publishing code.

    Free

    Security scanned
    Publish Pre-flight: make it public without leaking your day

    Publish Pre-flight: make it public without leaking your day

    Example session with this skill installed

    I'm ready to open-source my project. Can you check if I've left any client names or API keys in the history? The client is 'AcmeCorp'.

    • Read your context and instructions
    • Compiled the publish pre-flight

    I have initialized the preflight environment and audited the repository history. I searched for 'AcmeCorp' and common secret patterns across all commits.

    The scan identified one active secret in the current files and two instances of the client name in the commit history.

    Pre-flight Audit Findings

    LocationTypeFindingStatus
    config/settings.json:12Secret[SECRET MASKED] (Potential API Key)Active
    commit 8f2a1b:README.mdDenylist[TERM MASKED] (AcmeCorp)In History
    commit 4c9d3e:docs/plan.mdDenylist[TERM MASKED] (AcmeCorp)In History

    Risk Assessment:
    The API key in settings.json is currently staged. The client name 'AcmeCorp' exists in two past commits, meaning it will be visible to anyone who clones the repo even if you delete the files now.

    Next steps

    1. Rotate the key: Revoke the API key found in settings.json immediately as it must be considered compromised.
    2. Scrub history: Use git filter-repo to remove the client name from the past commits identified above.
    3. Install hooks: Run python scripts/preflight.py install to prevent these terms from being committed again.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Before you open-source a side project, check that it doesn’t carry your employer’s name, a client, a colleague’s email, a personal folder path or a live API key. Check the history too: Git keeps the key you “removed” three commits ago in every clone.

    Publish Pre-flight checks the files and the full history against your private denylist and a set of secret patterns, with masked output that’s safe to share. It then installs pre-commit, commit-msg and pre-push hooks so nothing private gets in later. It fails closed: if it can’t check, Git refuses. The skill guides your agent through writing the denylist, sorting the findings (rotate a leaked key first, then rewrite history), and handling false positives.

    Python 3.10+, nothing to install. Includes a demo repo with planted leaks, a triage guide, a hooks and CI guide, and 28 tests. MIT licensed.

    Public Git repositories often leak sensitive information through commit history even after the files are fixed. This skill audits your local repository and its entire history for employer names, internal codenames, and secrets before you publish or open-source the code.

    What it does

    • Denylist filtering blocks specific strings like client names, employer brands, and internal project codes.
    • Secret detection identifies high-risk patterns including API keys, private keys, and passwords.
    • Full history scanning inspects every past commit to ensure removed secrets do not remain in the Git history.
    • Hook installation configures pre-commit and pre-push hooks to prevent future leaks during development.
    • Path auditing flags personal home folder paths and outside email addresses that should stay private.

    How it works

    1. Initialize the environment to create the private denylist and local configuration files.
    2. Populate the denylist with sensitive terms and paths that must never appear in the codebase.
    3. Run the file scan and history scan to identify existing leaks across all branches and commits.
    4. Triage findings by rotating secrets, rewriting history, or marking false positives using specific comment tags.
    5. Install Git hooks to automate these checks for every future commit and push.

    Frameworks & tools

    This skill utilizes a Python 3.10+ script and Git 2.31+. It operates directly on the local filesystem and Git object database without external dependencies or cloud processing.

    Why this beats prompting it yourself

    LLMs often miss secrets buried in binary files or deep within Git history. This tool combines deterministic scanning of the entire Git object tree with a private, local-only denylist that ensures your sensitive data never leaves your machine.

    Use cases

    • Auditing a private work project before moving it to a public GitHub repository.
    • Cleaning up internal documentation that might contain colleague names or private email addresses.
    • Securing a side project that was developed in a directory containing company assets.
    • Hardening a repository against accidental API key commits via automated hooks.

    Known limitations

    Does not scan the content of images, PDFs, or other binary files. Cannot audit GitHub issues, Pull Request comments, or repository settings. Requires manual rotation of leaked keys.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.

    Does not scan binary files like images or PDFs. Cannot check data outside the local Git repo, such as GitHub issues or PR comments. Requires manual revocation of leaked secrets.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 11 days ago

    • Free with an Agensi account

    Software
    PythonGit

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Environment Variables
    Listed11 days ago

    What's inside

    Frequently Asked Questions

    Unlike a simple file scan, this skill audits the entire Git object tree, ensuring that sensitive data deleted in recent commits isn't still hiding in your project's hidden history.

    The skill is compatible with any AI agent or environment running Python 3.10+ and Git 2.31+, operating entirely on your local filesystem for maximum privacy.

    No, all scanning and denylist filtering happen locally on your machine; your sensitive terms and codebase are never uploaded to the cloud or sent to an LLM.

    You receive the core Python auditing script, a template for your private denylist, and the automation logic for installing pre-commit and pre-push Git hooks.

    The skill flags the location of the leak, but for security reasons, you must manually rotate the compromised keys and use standard Git tools to rewrite history if a leak is found.

    Yes, once configured, the skill can install hooks that automatically block commits or pushes if new secrets or denylist terms are detected, preventing future leaks before they happen.

    New in Agents & Orchestration

    Free