What problem does it solve? When a dependency, tool, or service turns out to be broken, undocumented, or wrong, reporting it well is hard: unverified claims waste maintainer time, duplicates add noise, and leaked hostnames or paths create privacy risk. This Skill turns a raw finding into a reproducible, privacy-checked GitHub issue that maintainers can act on. ## Core Features & Use Cases - Verification workflow: Reproduces every claim, searches existing issues (open and closed) with gh search issues, and reads the project's CONTRIBUTING and issue templates before drafting. - Minimal reproduction and redaction: Drills the bug down to a minimal case built in an empty directory, and strips hostnames, tokens, paths, and other identifying details from the report body. - Structured drafting and gated filing: Shapes the report around expected-vs-actual with evidence in <details> blocks, adds a one-line AI-disclaimer, renders a preview via the GitHub markdown API, and files with gh issue create only after explicit user approval. - Use Case: You discover a CLI tool silently drops data on a specific flag. The Skill verifies the behavior on a stated version, confirms no duplicate exists, drafts the issue in a subagent, shows you the rendered preview, and files it after your yes. ## Quick Start Report the bug I just found in the upstream repo to its maintainers as a GitHub issue.
