---
name: graph-search
description: Searches the code knowledge graph by function, class, or type using FTS5 full-text search. Use when locating code entities by name or qualified path.
model_hint: standard
---

# Search Code Knowledge Graph

Search `.gauntlet/graph.db` for code entities by name.

## When NOT To Use

- The graph is missing or stale (use `gauntlet:graph-build`)
- Scoring the impact of a change (use `pensive:blast-radius`)

## Steps

1. **Accept query**: Get the search term from the user.

2. **Run the query script**:
   ```bash
   python3 ${CLAUDE_PLUGIN_ROOT}/scripts/graph_query.py \
       --action search --query "<term>" --limit 20
   ```

   Optional filters:
   - `--kind Function` to search only functions
   - `--kind Class` to search only classes

3. **Display results**: Show qualified name, file path,
   line numbers, and relevance score for each match.

4. **Offer to read**: Ask if the user wants to read the
   top result's source file.

## Query Intelligence

The search engine detects query patterns:

- **PascalCase** (e.g., `UserService`): boosts Class
  and Type results
- **snake_case** (e.g., `get_users`): boosts Function
  results
- **Dotted path** (e.g., `app.models.User`): boosts
  qualified name matches

## Prerequisites

The graph must be built first. If `.gauntlet/graph.db`
does not exist, suggest running the `graph-build` skill.

## Exit Criteria

- [ ] If `.gauntlet/graph.db` does not exist, the skill surfaces
  the missing-prerequisite error and suggests `gauntlet:graph-build`
  rather than failing silently
- [ ] Results display qualified name, file path, line numbers, and
  relevance score for each match (up to `--limit` results)
- [ ] Query intelligence applies correct boost: PascalCase boosts
  Class/Type results, snake_case boosts Function results, dotted
  path boosts qualified name matches
- [ ] User is offered to read the top result's source file after
  results are shown
- [ ] A `count: 0` result is reported as "not in the graph" only after
  `graph_query.py --action status` shows `node_count` above zero. An
  unbuilt graph, or one built without tree-sitter, answers every
  query with zero, and that zero says nothing about the codebase