What problem does it solve?
When a library change breaks consumers, the failure often shows up only at integration time, so QA needs a fast, reproducible way to verify the library’s actual runtime behavior against the expected spec.
Core Features & Use Cases
- Behavioral verification via scratch consumers: Generates a small one-shot program that imports the library, calls the API, and emits structured (JSON) observations for reliable parsing.
- Change-focused scoping: Uses the recent changeset (git diff/log) to determine which exported symbols/types/behaviors to exercise when scope isn’t explicitly provided.
- Language-appropriate execution: Supports Go, Ruby (gem vs Rails), and TypeScript with runner-specific gotchas and output discipline.
- Evidence-driven verdicts: Captures stdout/stderr/exit codes, boot transcripts, parsed JSON envelopes, and a persisted verdict for APPROVE/REJECT/ESCALATE.
Quick Start
Use the library-qa skill to probe the changed exported surface by running a scratch consumer that calls the library API, prints structured JSON output, and produces a verdict from the parsed envelope.