feat: expand stub task descriptions into detailed agent prompts

[marcus]

Summary

• Expanded all 52 stub task descriptions in the registry from vague one-liners into detailed agent prompts with methodology, scope, and output format sections
• Preserved the 3 existing detailed descriptions (TaskSkillGroom, TaskTDReview, TaskPIIScanner) unchanged
• TaskGroomer now has a self-referential description enabling autonomous audit of task description quality

Details

Each expanded description follows the pattern established by the existing detailed descriptions and includes:

• WHAT TO LOOK FOR / WHAT TO EXAMINE — specific things the agent should check
• METHODOLOGY — numbered steps for the agent to follow
• OUTPUT FORMAT — expected fields and structure for findings
• SCOPE — boundaries on what not to do (where applicable)

Categories covered

Category	Task count	Description
PR (Category 1)	13	10 expanded, 3 already detailed
Analysis (Category 2)	18	17 expanded, 1 already detailed
Options (Category 3)	13	all 13 expanded
Safe (Category 4)	5	all 5 expanded
Map (Category 5)	7	all 7 expanded
Emergency (Category 6)	3	all 3 expanded

Test plan

• go vet ./internal/tasks/... passes
• go build ./internal/tasks/... passes
• go test ./internal/tasks/... passes
• Pre-commit hooks pass (gofmt, go vet, go build)

Nightshift-Task: task-groomer
Nightshift-Ref: https://github.com/marcus/nightshift

🤖 Generated with Claude Code

---

Automated by nightshift