Skip to content

Conversation

@huntclubhero
Copy link

Summary

Adds mandatory user confirmation before creating any skill, giving users full control over their skill library.

Builds on PR #7 (token efficiency refactor)

Changes

SKILL.md Updates

  1. New "User Confirmation Required" section - Makes explicit that skills are never created without approval

  2. Step 5: Request User Confirmation - Before saving, must present:

    I identified extractable knowledge:
    
    **Topic**: [brief description]
    **Proposed skill**: `[skill-name]`
    **Why**: [1-2 sentence justification]
    
    Would you like me to create this skill?
    
  3. Retrospective Mode - Now includes "Confirm" step before extraction

  4. Updated description - Metadata now states "Always asks for user confirmation"

  5. Version bump - 4.0.0 → 4.1.0

Rationale

The original design triggered skill extraction automatically after certain tasks. This could lead to:

  • Unwanted skills cluttering the library
  • Skills created for one-off solutions
  • Loss of user agency

With this change, users decide what gets saved. The skill still identifies candidates and does the work of structuring them, but the final decision is always the user's.

Test plan

  • Verify skill prompts for confirmation before creating
  • Confirm "No" response prevents skill creation
  • Test /claudeception retrospective asks which candidates to create
  • Verify description mentions confirmation behavior

🤖 Generated with Claude Code

huntclubhero and others added 2 commits January 19, 2026 21:56
- Reduce SKILL.md from ~3,500 words to ~1,800 words (~48% reduction)
- Extract detailed research workflow to resources/research-workflow.md
- Extract quality gates, anti-patterns, lifecycle to resources/quality-guidelines.md
- Use tables for compact information presentation
- Add clear references to existing examples/ directory
- Follows Claude Code skill development best practices for progressive disclosure

The refactored structure loads core instructions immediately while deferring
detailed reference material until needed, reducing context window usage.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add "User Confirmation Required" section making explicit approval mandatory
- Update extraction process: Step 5 now requests user confirmation
- Update Retrospective Mode to include "Confirm" step before extraction
- Update description to clarify skill always asks before creating
- Bump version to 4.1.0

Users now maintain full control over their skill library - no skills are
created without explicit approval.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant