Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
athola avatar

Graph Search

  • 74 installs
  • 325 repo stars
  • Updated August 2, 2026
  • athola/claude-night-market

Find functions, classes, and types in your repo via the Gauntlet FTS5 knowledge graph instead of blind grep or manual tree walks.

About

Graph Search is an agent skill for solo builders using Claude Night Market’s Gauntlet graph. When you know a symbol name but not its file, the skill accepts a query, invokes `graph_query.py search`, and surfaces ranked entities from the local SQLite knowledge graph. Pattern-aware ranking reduces noise compared to plain text search. Results include enough location metadata for your agent to jump straight into edits or reviews, with an explicit handoff to read the top file. The workflow assumes the graph was built beforehand; missing `.gauntlet/graph.db` triggers a clear prerequisite path to graph-build. Ideal for medium and large repos where agents otherwise burn tokens listing directories. Complexity is beginner-friendly once the graph exists; building the graph is the heavier one-time step.

  • FTS5 search on `.gauntlet/graph.db` with configurable `--limit` (default 20)
  • Optional `--kind Function` or `--kind Class` filters
  • Query intelligence: PascalCase boosts Class/Type, snake_case boosts Function, dotted paths boost qualified names
  • Returns qualified name, file path, line range, and relevance score
  • Offers to open top hit source; suggests graph-build if the DB is missing

Graph Search by the numbers

  • 74 all-time installs (skills.sh)
  • Ranked #5,535 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
  • Security screen: LOW risk (skills.sh audit)
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/athola/claude-night-market --skill graph-search

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs74
repo stars325
Security audit3 / 3 scanners passed
Last updatedAugust 2, 2026
Repositoryathola/claude-night-market

What it does

Find functions, classes, and types in your repo via the Gauntlet FTS5 knowledge graph instead of blind grep or manual tree walks.

Files

SKILL.mdMarkdownGitHub ↗

Search Code Knowledge Graph

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

Steps

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

2. Run the query script:

   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

Related skills

FAQ

Is Graph Search safe to install?

skills.sh reports 3 of 3 security scanners passed. Review the Security Audits panel on this page before installing in production.

AI & Agent Buildingintegrationsbackend

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.