Capture, consult, guard, scaffold, maintain, health-check, eval, and language-align project lore — the implicit knowledge (business rules, API maps, pitfalls, decision rationale) that code, types, and git history can't reveal.
8 skills · hooks
Visual regression testing — compare Figma designs against live web pages with pixel-level diff and AI vision.
2 skills · 1 command · 1 agent
Web search through Jev Search: answers Claude Code's built-in WebSearch with Jev-ranked results (sources and time window chosen by Jev) and falls back to the built-in tool when Jev cannot answer. Adds the jev_search MCP tool and a skill. Requires CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.
1 skill · 1 MCP server
Code review across standards, spec, and bugs in parallel sub-agents; important findings are confirmed by independent votes. Read-only; never modifies code.
1 skill
Test review doctor — diagnose an existing test file for blind spots and output a self-contained HTML report. Read-only; never modifies code.
1 skill
Review every AI session (Claude Code + Codex) of one or more projects and get an interactive HTML report: skill ranking, your real workflow as diagrams, how you talk to AI when building and debugging, and repeated routines worth turning into custom skills with a user/project scope call.
1 skill