Agent skill
web-rag
(ePost) Use when searching web codebase for components, design tokens, Next.js patterns, or any luz_next implementation via vector search
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/web-rag
Metadata
Additional technical details for this skill
- keywords
-
rag vector-search klara-theme luz-next web components design-tokens patterns implementation
- triggers
-
[ "search components", "find component", "klara-theme", "design token", "luz_next", "web pattern", "implementation" ] - platforms
-
[ "web" ] - connections
-
{ "enhances": [ "web-frontend" ] } - agent affinity
-
[ "epost-muji", "epost-fullstack-developer", "epost-planner", "epost-researcher" ]
SKILL.md
Web RAG Skill
Purpose
Vector search for the web codebase. Semantic search across luz_next application code, klara-theme component libraries, design tokens, and implementation patterns via MCP tools.
MCP Tools
| Tool | Purpose |
|---|---|
query |
Semantic search — returns code chunks with relevance scores |
status |
Health check, indexed modules, document counts |
catalog |
List available components/hooks/styles per module (no vector search) |
navigate |
O(1) shortcut to known code locations |
expansions |
Component alias mappings + synonym groups for query planning |
Tool parameters are self-describing — check inputSchema for current options.
Discovery Protocol
Before querying, discover what's indexed:
- Call
status— returns available modules, document counts, health - Call
catalog(optional) — lists components/hooks/styles for discovery - Call
expansions(once per session) — get component aliases and synonym groups
Use discovered values for module, filters.scope, and filters.component params.
When to Use
| Scenario | RAG | Grep/Glob | Context7 |
|---|---|---|---|
| Find component by concept | yes | if know filename | no |
| Search design tokens | yes | yes | no |
| Library API docs | no | no | yes |
| Pattern across codebase | yes | exact match only | no |
| Discover available components | catalog | no | no |
Query Strategy
- Known component or file? ->
query+ component filter (get canonical name fromexpansions) - Known topic area? ->
query+ topic/file_type filters (discover valid values viastatus) - Conceptual/cross-cutting? -> Smart query with HyDE (see
references/smart-query.md) - < 3 results or low scores? -> Broaden: remove filters, try synonyms, alternate casing
- Still nothing? -> Fall through to Grep/Glob (codebase search)
Rules
- Start broad, refine with filters — not longer queries
- Natural language works best — "button with loading state" over "btn loading"
- Low relevance scores suggest rephrasing needed
- If server offline, fall back to Grep/Glob
stale_sidecar: truein results — trust code chunks, ignore metadata fields- Do NOT generate synonym variants — server handles expansion automatically
Integration
Priority level 2 in knowledge-retrieval chain:
docs/files -> 2. RAG -> 3. Skills/codebase -> 4. Context7
Related Skills
knowledge-retrieval— Orchestrates source priorityweb-frontend— Frontend patterns (enhanced by this skill)docs-seeker— External documentation lookup
References
references/smart-query.md— HyDE + multi-query retrieval strategyreferences/sidecar-workflow.md— AI-generated metadata enrichmentreferences/component-mappings.md— Get canonical names viaexpansionsMCP toolreferences/synonym-groups.md— Get synonym groups viaexpansionsMCP tool
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?