Skip to content

feat: add /ai-hybrid skill — AI-human collaboration architect#158

Open
HMAKT99 wants to merge 1 commit intogarrytan:mainfrom
HMAKT99:arun/ai-hybrid-skill
Open

feat: add /ai-hybrid skill — AI-human collaboration architect#158
HMAKT99 wants to merge 1 commit intogarrytan:mainfrom
HMAKT99:arun/ai-hybrid-skill

Conversation

@HMAKT99
Copy link

@HMAKT99 HMAKT99 commented Mar 18, 2026

47% of your commits are AI-assisted. Is that making you faster or sloppier?

Every team uses Claude Code, Copilot, Cursor — but nobody measures the impact. Are AI-assisted commits higher quality or lower? Which tasks benefit from AI and which don't? What's the actual ROI?

What /ai-hybrid does

You:   /ai-hybrid --metrics

Claude: AI PRODUCTIVITY DASHBOARD
        Metric                      Before AI    After AI    Change
        Commits/developer/week      12           22          +83%
        Bug fix follow-ups          3/week       5/week      +67% ←
        Test coverage trend         78%          72%         -6%  ←

        TASK CLASSIFICATION:
        FULLY AUTOMATE:     Boilerplate, test writing, docs
        HUMAN-IN-THE-LOOP:  Feature implementation, refactoring
        HUMAN-LED:          Architecture decisions, security
        KEEP HUMAN:         Compliance, incident response

        AI CODE QUALITY AUDIT:
        Common smell: AI over-engineers — adds unnecessary abstractions
        Common smell: AI misses edge cases — happy path only

Analyzes AI usage from git history (Co-Authored-By patterns). Classifies tasks by optimal human-AI split. Audits AI-generated code quality. Designs optimized workflows with time-savings estimates.

Only .tmpl committed — bun run gen:skill-docs generates the rest.

Test plan

  • .tmpl follows template pipeline
  • Registered in gen-skill-docs.ts, skill-check.ts, test files

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