feedback-intake

# Nimbalyst Feedback Skill This skill is the entry point for the Nimbalyst in-app feedback flow. It activates when: - The user runs the `/feedback:bug-report` slash command - The user runs the `/feedback:feature-request` slash command - The user opens a session with `sessionType: 'bug-report'` or `sessionType: 'feature-request'` ## Behavior Read the matching command file: - For bug reports: `commands/bug-report.md` in this plugin - For feature requests: `commands/feature-request.md` in this plugin Both commands share these constraints: 1. **The user is the gate.** Never call `mcp__nimbalyst-situational__feedback_open_github_issue` without explicit user approval of the final title and body. 2. **Anonymize aggressively.** Run anything that includes file paths, log content, environment strings, or workspace identifiers through `mcp__nimbalyst-situational__feedback_anonymize_text` first, then read the output and redact anything else that looks sensitive. 3. **Honor the intake flags.** Bug-report sessions include `Log gathering: allowed` or `Log gathering: not allowed`; feature-request sessions may include `UX mockup: requested` or `UX mockup: not requested`. Only bug reports may read logs, and only when log gathering is allowed. 4. **Handle long reports.** When the drafted body exceeds the URL safe length, the URL builder returns `reason: 'too-long'`. Fall back to: show the body in chat, open a no-body URL, tell the user to paste. 5. **Target repo and templates.** Bug repor...