Code Factory Audit

    1

    Check AI-written code locally before you merge it: finds tests that cannot fail and bounded Python security issues, with an evidence trail.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Code Factory Audit

    Example session with this skill installed

    Run Code Factory Audit on this repo. A contractor added a pricing module and I want to know if its tests can actually fail before I merge.

    • Read your context and instructions
    • Compiled the code factory audit

    I ran factory scan --root . --json on the repository. Real output from a two-file sample project, summarized:

    Scan summary

    FieldValue
    Files discovered2 (both Python)
    Overall stateINCOMPLETE
    Static securityFINDINGS (1)

    Finding

    CodeSeverityLocationMessage
    QUALITY_HOLLOW_TESTMEDIUMtests/test_pricing.py:3Test has no local non-vacuous assertion. Add an independent expected outcome and a negative control.

    The test test_apply_discount computes a result and then asserts True. It passes whatever apply_discount returns, so its green status says nothing about the discount logic.

    Not checked

    The overall state is INCOMPLETE, not clean. The scan reports that no project coverage receipt is bound and no signed native worker is configured, so those evidence lanes did not run.

    Next steps

    1. Replace assert True with assert result == 90.0, and add a case that fails when the discount is wrong.
    2. Re-run the project tests and factory scan --root . to confirm the finding clears.
    3. For a fuller review, run factory scan --root . --deep and read the unmeasured lanes before deciding to merge.

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

    About this skill

    Check code you did not write before it reaches your main branch. Green tests prove nothing if a test cannot fail. This skill runs Code Factory, an open-source local scanner, and walks your agent through a scan, repair and re-verify loop with a written evidence trail.

    What it does

    • Scans a local repository (read-only). Nothing is executed, modified or uploaded.
    • Flags tests with no real assertion (QUALITY_HOLLOW_TEST) and bounded Python security findings.
    • Drives one-finding-at-a-time repairs, then re-runs your tests and the scan.
    • Reports what was never checked. Missing evidence is shown as INCOMPLETE, never as a pass.
    • Optional --deep mode adds architecture, requirements, workflow, runtime coverage and dependency lanes.

    What you need

    Python 3.10+ and pip install factoryline-code-factory (free, MIT OR Apache-2.0). Source: https://github.com/zrk222/code-factory

    Honest limits

    • Python source gets security and test-quality analysis. Other languages are inventoried by file type only, with no security verdict.
    • It is a bounded static review, not a penetration test. Receipts support review; they do not certify code or approve a release.
    • It complements pytest, mutation testing, CodeQL, Semgrep and CI. It does not replace them.

    Good for

    • Reviewing code an AI agent or contractor wrote
    • Checking a PR before merge
    • Finding green tests that assert nothing

    Launch offer

    Free through December 15, 2026.

    Requires local source checkout. Python 3.10+ only. Deep worker configurations (CodeQL/Semgrep) require manual setup. Python is analyzed; other languages are inventoried only.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 25 days ago

    • Free with an Agensi account

    Requires

    Python 3.10+
    pip install factoryline-code-factory
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Listed25 days ago

    What's inside

    Frequently Asked Questions

    It checks the source locally for tests that cannot fail (such as a test that only asserts True) and for bounded Python security findings, then reports what was never checked. Missing evidence is shown as INCOMPLETE, never as a pass.

    Python gets the security and test-quality analysis, and the repair loop applies to those findings. TypeScript/JSX, Go, Rust, Java, C/C++, C#, Ruby, PHP and Swift are inventoried by file type only, with no security verdict.

    A prompt-only review has no evidence behind it. Code Factory runs a deterministic local scan and records a receipt tied to a source hash, so you can see what was checked and what was not. It supports your review; it does not certify the code.

    You will receive the 'code factory audit' skill definition, access to the `factoryline-code-factory` package requirements, and documentation for integrating the scan-repair-verify loop into your workflow.

    Yes, it is designed to integrate with your existing environment, working alongside tools like pytest, CodeQL, and Semgrep to enhance your current CI/CD or local development pipeline.

    New in Code Quality & Review