Auto-loads .vibe/ project context for commands and agents. Triggers on code questions, implementation, reviews, planning. Provides role-based context loading and API lookup chain.
From vibenpx claudepluginhub jwlutz/claude_code_frameworkThis skill uses the workspace's default tool permissions.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Guides idea refinement into designs: explores context, asks questions one-by-one, proposes approaches, presents sections for approval, writes/review specs before coding.
Check if .vibe/ exists. If not: "No vibe context found. Run /vibe:init to analyze this codebase."
Load files based on caller role:
| Role | Loads |
|---|---|
| explorer | understanding.md (full), risks.md, bugs.md, current.md, docs index |
| engineer | understanding.md (patterns, components, tests), current.md, decisions.md (recent 5) |
| reviewer | understanding.md (patterns, tests), risks.md (baseline), bugs.md |
| tester | understanding.md (tests section only), bugs.md (for regression checks) |
| general | understanding.md, current.md |
Read docs index from understanding.md. If current task relates to an indexed doc, load that doc on-demand from .vibe/docs/.
Before any agent uses an unfamiliar API, run this chain. Never guess. Always verify.
resolve-library-id then query-docs. Use result. State: "Verified via Context7."NEEDS_CONTEXT to controller. Do not proceed with unverified API.Agents must state which step provided their information.
When an engineer or implementation phase completes (not per-file, once per phase):
debug/When writing to .vibe/ files, follow these rules: