Debugging preflight

    1

    Standardized diagnostic protocol to isolate setup blockers from application bugs before editing code.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Debugging preflight

    Example session with this skill installed

    My local 'npm test' is failing with a connection error but I haven't changed the database logic. Check the preflight for this.

    • Read your context and instructions
    • Compiled the debugging preflight

    I have performed a debugging preflight check on the reported test failure. I compared the repository requirements against the current environment state and attempted a localized reproduction.

    Preflight Diagnostic Report

    CategoryDetails
    Expected Behaviornpm test should connect to the local Postgres instance and execute the suite.
    Actual BehaviorTest runner terminates during the beforeAll hook with ECONNREFUSED.
    Reproduction ProcedureChecked docker ps for active containers and inspected .env metadata.
    Failure PhaseSetup/Initialization (Database Connection).
    Setup EvidenceFound .env.example requires DB_URL, but local .env is missing the port assignment.
    StatusSetup Blocker Detected.
    UnverifiedApplication logic for the search and update modules remains unverified.

    Observations

    • Supplied: User reported connection error.
    • Reproduced: Confirmed ECONNREFUSED by running the specific failing test file.
    • Inference: The database service is likely not running or the port mapping in .env is incorrect.

    Next steps

    1. Verify if the Docker container for the database is running using docker-compose up -d.
    2. Sync your local .env with the DB_URL specified in .env.example.
    3. Rerun the single failing test to confirm the connection is established.

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

    About this skill

    Debugging a reported failure often leads to wasted tokens editing application code when the root cause was actually a missing environment variable or a misconfigured runtime. This skill enforces a rigorous pre-diagnostic protocol to separate setup issues from genuine application bugs before any code is touched. It prioritizes read-only inspection and minimal intrusion to protect your development environment from side-effect-heavy troubleshooting.

    What it does

    • Setup verification checks documented runtimes, dependencies, and service availability against project requirements.
    • Smallest-path reproduction identifies the minimal set of inputs and procedures needed to trigger a reported discrepancy.
    • Phase identification isolates whether a failure occurs during initialization, collection, or execution.
    • Evidence categorization distinguishes between user-supplied logs, static code inference, and fresh local observations.
    • Blocker reporting stops diagnostic runs immediately when a missing prerequisite is found, preventing false positives in bug reporting.

    How it works

    1. Instruction Audit: The agent scans the repository for runtime, package-manager, and service requirements.
    2. Minimal Intrusion Check: It performs the least invasive checks to distinguish environment setup from application logic.
    3. Reproduction Attempt: It attempts to reach the reported failure using safe, nonproduction commands.
    4. Outcome Categorization: It produces a report documenting expected vs. actual behavior and the specific failure phase.

    Frameworks & tools

    This skill is framework-agnostic and works with any stack utilizing documented setup files, including package.json, requirements.txt, .env.example, and standard test runners or CLI tools.

    Why this beats prompting it yourself

    Standard LLM prompts often jump straight to refactoring code when shown an error log. This skill forces the agent to treat logs as leads rather than proof, preventing "fix-loops" caused by underlying environment issues that code edits cannot solve.

    Use cases

    • Diagnose why a CI pipeline is failing locally while passing for others.
    • Verify if a "bug" is actually just a missing .env key or wrong Node version.
    • Establish a clean reproduction case for a complex application crash.
    • Audit service availability and configuration metadata without leaking secrets.

    Known limitations

    Does not perform broad package installations or start services automatically. Requires user authorization for any isolated nonproduction checks that might involve side effects.

    Diagnosis is read-only by default. Does not print secrets or dump full environments. Stops at the first setup blocker found.

    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

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 16 days ago

    • Free with an Agensi account

    Permissions

    Read Files
    Listed16 days ago

    What's inside

    Frequently Asked Questions

    This skill prevents 'fix-loops' by forcing the agent to verify environment variables, runtimes, and dependencies before attempting any code changes, ensuring you don't waste time refactoring logic for an configuration-based error.

    The skill is framework-agnostic and works with any stack that uses standard configuration files like package.json, requirements.txt, or .env templates to define project requirements.

    No, the skill prioritizes read-only inspection and minimal intrusion; it will identify missing prerequisites and report them as blockers rather than attempting to install software or modify your system.

    You receive the standardized diagnostic protocol, the reproduction logic, and the reporting templates designed to integrate with your existing AI agent's workflow.

    This skill is designed for agents with file-system and terminal access, as it requires the ability to scan repository files and run non-intrusive diagnostic commands.

    Yes, the protocol includes a specific 'Smallest-path reproduction' phase designed to isolate the minimal conditions needed to trigger a failure, making it ideal for documenting complex crashes.

    New in Testing & Debugging