Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    ai Trace Runner

    2

    Transform ambiguous AI tasks into auditable execution traces with verified evidence and AI-smell detection.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    ai-trace-runner

    ai-trace-runner

    Example session with this skill installed

    Trace a run to update the API routes and verify test coverage. Stop if you hit blockers. Clean up the final summary for AI-smell.

    • Read your context and instructions
    • Compiled the ai-trace-runner

    [Verification Report]

    • Criterion: 'Update API routes' -> VERIFIED (See diff in Evidence Bundle)
    • Criterion: 'Test coverage > 80%' -> BLOCKED (Missing jest config)

    [AI-Smell Report]
    Score: 4/5 (Repetitive hedging detected)
    Cleanup: Removed "As an AI..." and "I have successfully..." phrases.

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

    About this skill

    What it does

    AI Trace Runner is a comprehensive execution wrapper for AI agents that turns ambiguous tasks into auditable, verifiable run records. It replaces basic prompting with a structured pipeline: creating a request contract, generating an execution plan, logging every material action, and assembling an evidence bundle. It ensures that every claim made by the AI is backed by a specific file path, command output, or observation.

    Why use this skill

    Standard AI agents often suffer from 'hallucinated confidence'—claiming a task is done when it actually failed or was bypassed. This skill solves the lack of transparency in agentic workflows by forcing explicit verification. It includes built-in "AI-smell" detection to strip away robotic fluff and project.yaml integration for rigorous dependency management. It is ideal for developers who need to debug complex agent behaviors or provide ironclad proof of work for code changes and system audits.

    Supported tools

    • Standard CLI and shell environments
    • Project management via project.yaml execution topologies
    • Any file-system or API-based tool within the agent's scope
    • Integrated human-in-the-loop gates for destructive actions

    • Trace quality depends on underlying tool output granularity.
    • High-latency for simple tasks due to multi-stage verification.

    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 6 months ago

    Requires

    GitHub
    Npmjs

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Allowed Hosts

    registry.npmjs.org
    github.com

    File Scopes

    examples/**
    src/**
    Listed6 months ago

    What's inside

    Frequently Asked Questions

    AI Trace Runner solves the 'black box' problem by forcing agents to document every step. It prevents hallucinated success by requiring verifiable evidence (like file outputs or terminal logs) for every claim made during a task.

    The skill is designed to work within standard CLI and shell environments. It integrates seamlessly with any agent capable of using file-system tools and is specifically optimized for workflows using `project.yaml` for dependency management.

    Every execution generates an 'Evidence Bundle' containing an execution plan, an audit log of all material actions, and specific file/command snapshots that prove the work was performed as described.

    Yes, the skill includes a 'Human-in-the-Loop' gate mechanism. This allows you to set manual approval checkpoints for destructive actions, ensuring the agent doesn't execute high-risk commands without oversight.

    'AI-smell' detection is a specialized filter that identifies and strips away generic, robotic filler text from agent responses. This ensures your logs contain only high-signal technical data and verifiable facts rather than conversational fluff.

    You will receive the core execution wrapper, the verification logic, and documentation for integrating the skill into your existing agent pipelines and `project.yaml` topologies.

    New in Agents & Orchestration