Merge pull request #739 from alirezarezvani/claude/andreessen-productivity-plugin-PlOzn

[alirezarezvani]

Summary

Checklist

• Target branch is dev (not main — PRs to main will be auto-closed)
• Skill has SKILL.md with valid YAML frontmatter (name, description, license)
• Scripts (if any) run with --help without errors
• No hardcoded API keys, tokens, or secrets
• No vendor-locked dependencies without open-source fallback
• Follows existing directory structure (domain/skill-name/SKILL.md)

Type of Change

• New skill
• Improvement to existing skill
• Bug fix
• Documentation
• Infrastructure / CI

Testing

[claude]

Code Review — PR #746

Summary: This is a merged dev → main PR that: (1) restores the .vibe/skills/claude-skills/skills-index.json after it was truncated to 6 entries by a previous --domain productivity sync, and (2) lands the Andreessen productivity plugin from PR #739.

---

What's good

• Clear root-cause commit message — the fix commit explains exactly what broke (--domain productivity clobbered the full index) and why the fix (full re-sync) works. Future maintainers will understand this immediately.
• Minimal blast radius — only the index JSON is touched. Symlink tree and all skill files are unaffected, which is the right scope.
• Fast recovery — the fix was identified and landed quickly.

---

Issues Found

1. total_skills: 323 is inconsistent with file content (data integrity)

The fix re-ran the sync without the --domain filter, adding 323 entries across 13 domains. However, the 6 productivity entries from the old index are preserved in the file as unchanged lines (visible in the diff as context, e.g. "path": "productivity/reflect"). The total_skills counter reflects only the new additions, not the combined total — so the field likely undercounts by the number of productivity skills in the old index.

The counter should match jq '[.domains[]] | add | length' skills-index.json.

2. Productivity domain skipped by the fix sync (14 → 13 domains)

The commit message claims "14 domains" but the added entries only cover 13 top-level domain prefixes:

business-growth, business-operations, c-level-advisor, commercial, engineering, engineering-team, finance, marketing-skill, marketing, product-team, project-management, ra-qm-team, research

productivity/ is absent from the new additions. This means the productivity domain's index entries are still from the old --domain productivity run — different provenance than the rest. The new andreessen skill from PR #739 was the whole point of that PR, but it only lands in the index if it was already in those 6 old entries.

Worth verifying: does andreessen actually appear in the current state of the file on disk?

3. Five skills have "description": ">" (sync script bug)

Five entries were indexed with a literal > as their description:

• engineering/command-guide
• engineering/ship-gate
• engineering/skill-security-auditor
• engineering/caveman
• engineering-team/epic-design

This looks like the sync script is picking up a markdown blockquote marker from a SKILL.md file instead of the actual description text. These should surface as failures in scripts/audit_skills.py. Worth investigating whether the description: YAML front-matter in those SKILL.md files has a malformed value.

---

Root Cause Remains Open

The underlying footgun is that sync-vibe-skills.py --domain X silently overwrites the full skills-index.json with only domain X, dropping all other domains. This will happen again. Consider either:

• Option A: Make --domain filter append/update rather than replace the full index.
• Option B: Add a guard that warns if total_skills drops by more than 10% vs the previous file.

---

Process

• PR description body is empty — future PRs should at minimum note what's being merged and how it was tested.
• PR checklist is entirely unchecked. For fix PRs, checking the applicable boxes (e.g. "no hardcoded secrets") signals the author reviewed the change, even if it's minor.

---

Overall: The fix is correct and necessary, and the commit message is exemplary. The issues above are follow-up items rather than blockers, but the total_skills count and missing productivity domain regeneration are worth a quick follow-up sync to make the index self-consistent.