- Home
- Skills
- Testing & Debugging
- FailFirst Lab — Bug-to-Test Regression Factory
More screenshots
Works with the AI Agents you already use
FailFirst Lab — Bug-to-Test Regression Factory
Convert a specific reproducible bug into an isolated failing test and evidence-backed fix verification.
$15
FailFirst Lab — Bug-to-Test Regression Factory
Example session with this skill installed
Run FailFirst Lab — Bug-to-Test Regression Factory as an evidence-first preflight. This is a representative, hypothetical test case: A cart total duplicates sales tax whenever a voucher is applied twice. The materials I have available are: bug report with observed and expected behavior; source repository and test framework; runtime logs or failing example. Tell me what is evidenced, what is missing, which claims must be held, the exact corrections, and a clean verification plan. Do not claim you opened a file or performed a live run unless one is actually supplied.
- Read your context and instructions
- Compiled the failfirst lab
REPRESENTATIVE EXAMPLE — NOT A LIVE RUN
PREFLIGHT: HOLD / VERIFY BEFORE ACTING. Task: Convert a specific reproducible bug into an isolated failing test and evidence-backed fix verification.
Scenario supplied by buyer: A cart total duplicates sales tax whenever a voucher is applied twice.
-
MATERIAL FINDING (HIGH): The user-provided issue description supplies a candidate reproduction, not a recorded failing run. The case describes a potential issue; no underlying account, file or production service has been accessed by this illustrative demo.
-
TRACEABILITY MATRIX
- Claimed behavior: A cart total duplicates sales tax whenever a voucher is applied twice.
- Critical evidence sources: bug report with observed and expected behavior, source repository and test framework, runtime logs or failing example.
- Status: NOT VERIFIED from primary artifacts in this example. A real invocation must attach the files or authorize a read-only source inspection and record exact references.
- CHECKS THE SKILL WILL RUN
- Reproduce the actual bug before changing code; mark BLOCKED if the issue cannot be observed.
- Identify a narrow deterministic fixture that fails on the old revision and expresses user behavior.
- Preserve existing framework conventions, mocking style and test naming; avoid snapshot-only false confidence.
- Apply the minimum authorized fix and verify fail-before/pass-after with unchanged assertions.
- Report trace, test path, commands, touched files and remaining untested cases without invented passes.
-
CORRECTION TO PROPOSE
Create two-voucher fixture, assert tax only once, then isolate the surcharge logic after reproducing. Do not present this as implemented until the seller/developer actually applies the change and reviews the result. -
REPRODUCTION / ACCEPTANCE TEST
Run the new test before the change (must fail) and after (must pass), plus full relevant suite. Capture the before/after witness (paths, commands, timestamp, observed outcome), distinguish a script signal from a working end-to-end product, and do not invent a PASS. -
SCOPE LIMIT
Requires runnable source and test environment for a verified patch; otherwise outputs a test plan only.
Final handoff: a clear evidence-linked issue list, copy-ready proposed changes where appropriate, a prioritized checklist, and a statement of exactly which tests were and were not run.
Connects securely to your tools. The creator never sees your data.
About this skill
FailFirst Lab — Bug-to-Test Regression Factory
Convert a specific reproducible bug into an isolated failing test and evidence-backed fix verification.
What you receive
- A ready-to-install skill with a repeatable, evidence-first procedure.
- A source-to-claim discrepancy matrix, severity tags and specific corrective actions.
- A clearly separated verification checklist and unknowns register.
- A representative worked example; it is not passed off as a live customer run.
Example issue
A cart total duplicates sales tax whenever a voucher is applied twice.
Trust boundary
Requires runnable source and test environment for a verified patch; otherwise outputs a test plan only.
No paid API subscription is bundled. The agent can inspect materials supplied by the buyer in its supported environment. Access to private services is never presumed.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.
Requires runnable source and test environment for a verified patch; otherwise outputs a test plan only.
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 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 8 days ago
Permissions
File Scopes
No production changes or account access without owner authorization.