- Home
- Skills
- Testing & Debugging
- Debugging preflight
Debugging preflight
Standardized diagnostic protocol to isolate setup blockers from application bugs before editing code.
Free
Works with the AI Agents you already use
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
| Category | Details |
|---|---|
| Expected Behavior | npm test should connect to the local Postgres instance and execute the suite. |
| Actual Behavior | Test runner terminates during the beforeAll hook with ECONNREFUSED. |
| Reproduction Procedure | Checked docker ps for active containers and inspected .env metadata. |
| Failure Phase | Setup/Initialization (Database Connection). |
| Setup Evidence | Found .env.example requires DB_URL, but local .env is missing the port assignment. |
| Status | Setup Blocker Detected. |
| Unverified | Application logic for the search and update modules remains unverified. |
Observations
- Supplied: User reported connection error.
- Reproduced: Confirmed
ECONNREFUSEDby running the specific failing test file. - Inference: The database service is likely not running or the port mapping in
.envis incorrect.
Next steps
- Verify if the Docker container for the database is running using
docker-compose up -d. - Sync your local
.envwith theDB_URLspecified in.env.example. - 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
- Instruction Audit: The agent scans the repository for runtime, package-manager, and service requirements.
- Minimal Intrusion Check: It performs the least invasive checks to distinguish environment setup from application logic.
- Reproduction Attempt: It attempts to reach the reported failure using safe, nonproduction commands.
- 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
.envkey 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
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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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