More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Ouroboros Evidence Guardian

    1

    Audit Git changes, CI claims and deterministic-code risks with revision-bound evidence. Separate verified facts from assumptions and get actionable merge-readiness reports without fabricated runtime success.

    $49

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Ouroboros Evidence Guardian

    Ouroboros Evidence Guardian

    Example session with this skill installed

    Review this Git repository for deterministic simulation risks and tell me whether a merge is justified. Record the exact HEAD SHA, inspect tracked source files, separate verified findings from heuristic warnings, and explicitly list any CI or production checks you cannot verify. Do not modify the repository.

    • Read your context and instructions
    • Compiled the ouroboros evidence guardian

    Summary
    Decision: BLOCKED pending independent CI and runtime verification. The static scan completed against a fictional sample checkout and flagged two possible deterministic-state risks. A match does not prove either call influences game state.
    Evidence scope
    Repository: illustrative sample project
    Revision: EXAMPLE_SHA (placeholder, not a real commit)
    Worktree: assumed clean for this demonstration
    Audit mode: read-only, local tracked-source heuristic
    Severity
    Surface
    Location
    Observation
    Status
    Medium
    Core
    src/simulation/tick.ts:14
    Date.now() candidate; inspect whether wall-clock time reaches the authoritative tick
    UNVERIFIED
    Medium
    Core
    src/world/spawn.ts:31
    Math.random() candidate; inspect whether a deterministic seed is required
    UNVERIFIED
    High
    CI
    No run receipt supplied
    Exact-revision test results unavailable
    BLOCKED
    High
    Runtime
    No runtime receipt supplied
    Live replay and tick invariants unverified
    BLOCKED
    Recommendations
    Trace the Date.now() call site to its consumers. If it feeds authoritative simulation state, replace it with the existing deterministic tick abstraction.
    Trace the random call. If it contributes to simulation output, use the system's reproducible seeded generator.
    Run deterministic replay tests against the exact commit under review, record command, environment, exit code and output fingerprint.
    Obtain an independent CI readback linked to that same SHA.
    Reassess the merge only when required checks and real runtime boundaries are supported by evidence.
    Limitations: This example is fictional and intentionally marked as such. No live repository, deployment, or CI was inspected for this listing demonstration.

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

    About this skill

    Stop treating a green badge as proof that software actually works. Ouroboros Evidence Guardian helps AI coding agents review Git repositories, pull requests, and release evidence with a consistent, revision-bound audit workflow. It records the exact Git SHA, inspects relevant source and checks, flags potential nondeterministic patterns, and separates verified observations from missing or blocked runtime proof. What you get A structured, reproducible read-only Git audit workflow. A bundled Python 3 standard-library scanner for tracked source files. Risk leads for Date.now(), Math.random() and related APIs. Evidence grading: VERIFIED / UNVERIFIED / FAILED / BLOCKED. Merge-readiness guidance that never treats static analysis alone as live runtime proof. A clear report format for findings, affected files, revision and next steps. Who it's for: AI-assisted developers, maintainers, technical leads, and teams using GitHub, GitLab or local Git repositories who want fewer unjustified "all good" claims. Important limitations: The included scanner is a heuristic, not a full static analyzer. CI, deployment and production evidence need access to the relevant systems. This product does not guarantee bug-free code, verify live production by itself, or automatically merge changes. No paid dependencies; Python 3 and Git required for the included scanner. Internet and account access are optional depending on which external evidence you ask the agent to verify.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.

    Requires Python 3 for the scanner script. Cannot prove absolute correctness, only identifies heuristic leads. Audit is blocked if Git metadata or repository access is unavailable.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 17 days ago

    Accounts
    GitHub

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Allowed Hosts

    api.github.com
    Listed17 days ago

    What's inside

    Frequently Asked Questions

    Traditional LLMs often trust PR descriptions; this skill ignores them and demands verifiable logs tied to a specific Git SHA, preventing "hallucinated passes" in your release cycle.

    The skill requires a Python 3 environment and Git access to your local or remote repository to execute the `audit_repo.py` scanner and generate evidence reports.

    You receive the core auditing scripts, the fail-closed logic prompt, and the JSON evidence schema designed to integrate with your existing AI agent's workflow.

    The skill is built for high-stakes environments like mission-critical systems or regulated software where you cannot afford 'floating' code claims that aren't tied to a specific revision.

    Yes, by generating standardized JSON evidence reports, it can be integrated into your CI/CD pipeline to flag unstable serialization or non-deterministic code before it reaches production.

    New in Code Quality & Review