More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Anti-Dark-Code: Understand and Verify Code

    2

    Make unfamiliar code easier to work with: map how it fits together, trace findings to evidence and check the changes you authorize.

    $29

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    About this skill

    Get a clearer picture of the code before deciding what to change.

    Anti-Dark-Code helps your AI assistant map an unfamiliar repository, investigate a problem and check a proposed repair. Findings point back to code and test evidence, so you can see what is known and what still needs checking.

    Optional bundle: Get both skills for $39 one-time. Each skill also works independently.

    What you can ask it to do

    • Explain entry points, responsibilities and dependencies.
    • Investigate a bug and trace the finding to its source.
    • Improve explanations and documentation while preserving behavior.
    • Choose proportionate checks, inspect test strength and report actual results.
    • Propose focused repairs and carry out the changes you authorize.

    Evidence you can use

    Reports distinguish observations, inferences and unknowns. They show which checks ran and what those results cover. A passing check supports the behavior it exercised; it is not a promise that every defect has been found. Source-bound records keep the reasoning available for later work.

    Your project, your decisions

    You choose the scope and approve consequential actions. Usage observation stays off unless you enable it. If enabled, it can read selected local host logs and keep counters and technical metadata, with no transcript bodies or uploads by default. Optional GitHub checks use the access you supply; managed-file changes and command execution follow your authorization.

    Works independently

    Use Anti-Dark-Code with your existing repository and documents. A one-off review needs no new planning document set or managed installation. Scaffold Kit is optional if you also want a structured plan for the next build.

    Getting started

    Use a compatible AI agent with access to the supporting files and the project material you authorize. Instructions can guide a manual review. Bundled helpers need Python 3.12 or newer and can run separately from an older-Python project. Git is needed for Git-based provenance. Your AI account and model/API usage are separate.

    Delivery and example

    Agensi handles checkout, access and delivery. The images show a fictional client-tracker Python example: five of seven checks pass before a proposed correction, and all seven pass in a separate corrected copy. The original is preserved. This is example logic, not a built application.

    Source and license

    Packaged from stable 2026.10.22-unified.23. Review the publicly available source and included FSL-1.1-MIT terms: https://github.com/LynxTWO/anti-dark-code-skill/releases/tag/v2026.10.22-unified.23

    Findings apply to the code examined and checks actually run. Runtime, file access and execution depend on your host and the project tools. Advanced managed migration uses the original release operator guide.

    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 any payment.

    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 25 days ago

    Requires

    Python 3.12+
    Git
    Project-specific tools
    Python

    Permissions

    Terminal / Shell
    Write Files
    Network Access
    Read Files
    Environment Variables

    Reads authorized project files. Writing and command execution are optional and require approval for their actual effects. Network/environment access is used by optional GitHub CI retrieval or approved project commands; GitHub is not required for a basic local review. Usage collection is off until explicit opt-in and may then read selected local host logs. example.invalid and h in the package are synthetic negative-test fixtures, not required external services. Scaffold Kit is optional.

    Listed1 month ago
    Updated18 days ago

    What's inside

    Frequently Asked Questions

    No. Anti-dark-code works independently with your existing repository, documentation and task scope. A one-off review does not require a Scaffold Kit plan or managed local installation. Scaffold Kit is an optional companion for planning intended work.

    A compatible AI agent with access to the supporting skill files and authorized project material. Bundled helpers require Python 3.12+; Git-based provenance features require Git. Manual inspection can proceed without helpers, with missing checks reported. Your own model/API usage and project tools are separate.

    Usage observation is off until you explicitly enable it. When enabled, it can read selected local host logs and retain usage counters and technical metadata. It does not retain transcript bodies or upload by default.

    No. An audit request does not authorize implementation. Commands and changes must match your approved scope and effects; protected changes require explicit owner approval. A configuration field or another agent cannot grant that permission.

    Results apply to the named scope and available evidence. The report separates observed facts, inferences and unknowns, and records which checks actually ran.

    New in Code Quality & Review