What problem does it solve?
This Skill captures a researcher's reaction when an advisor's response "felt off", clarifies the complaint in one brief exchange, classifies the failure mode, and produces a privacy-safe, diagnostic-quality issue report that maintainers can act on without exposing the researcher's work.
Core Features & Use Cases
- Conversational Clarification: Warmly acknowledge the user's reaction and ask at most one follow-up question to clarify tone, substance, or accuracy problems.
- Issue Classification: Internally classify failures into types such as voice, accuracy, blind-spots, caricature, disagreement, coverage, wrong-lens, or other and assign severity levels.
- Diagnostic Report Builder: Construct a reproducible, privacy-preserving issue body with Persona, Issue Type, Mode, a generic Question Type, failure description, recommended fixes, and an anonymous researcher ID.
- Submission & Fallbacks: Attempt direct submission to an HTTP feedback endpoint, fall back to the GitHub CLI if available, and save to a local pending queue as a last resort.
- Use Case: A researcher says "that advice sounded generic" and the Skill confirms once, diagnoses which persona commitments failed, files a ticket for maintainers, and confirms completion to the user.
Quick Start
Say that an advisor's response "felt off" and briefly describe what was wrong so the assistant can file privacy-safe diagnostic feedback.