Enforced methodology for creating production-grade AI agent skills. Mandates skill-creator for drafting and skill-evaluate for optimization, with A/B testing at each gate. No shortcuts, no bypass.
★not rated 21 todayA
tokens not measured
originalMIT
Pragmatic AI coach: diagnoses your real need, gives a clear verdict (no AI needed / yes here's how / bad idea here's why), and redirects to the right HORKA plugin.
★not rated 21 todayA
tokens not measured
originalMIT
Global index of your Claude Code projects. Disambiguates generic commands ("write the changelog") by resolving a name/alias to a path, knowing which folder owns a skill/command, detecting parent workspaces (monorepos), and creating + indexing new projects in the right place.
★not rated 21 todayA
tokens not measured
originalMIT
Complete code review toolkit: Tech Lead PR review agent + antagonist reviewer + review guard hook (blocks push to the default branch, with a per-machine allowlist). Catches quality, security, and architecture issues before they reach production.
★not rated 21
changed todayA
tokens not measured
originalMIT
Evaluate skills against Anthropic's official best practices guide with scoring, fixes, and before/after debrief.
★not rated 21 todayA
tokens not measured
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: