Overview
Code Graph Query Skill is the graph-query skill from AI Maestro, the MIT-licensed dashboard by 23blocks for running many coding agents with persistent memory, messaging, and multi-machine support. AI Maestro indexes your codebase into a code graph (also shown as an interactive visualization in the dashboard). This skill makes the agent query that graph before modifying any component, to find callers, dependencies, and the blast radius of a change, so it does not break code it has not read.
The rule it enforces is Read file, query graph, then proceed: query right after reading a code file or finding it through search, before making any modification, and include the findings in the analysis. The skill is published in ai-maestro-plugins and requires AI Maestro with Bash access; commands auto-detect the agent ID from the tmux session.
Commands
| Command | Purpose |
|---|---|
graph-describe.sh <name> | Describe a component or function |
graph-find-callers.sh <function> | Functions that call this function |
graph-find-callees.sh <function> | Functions this function calls |
graph-find-related.sh <component> | Related components (extends, includes, and so on) |
graph-find-by-type.sh <type> | All components of a type: model, serializer, controller, service, job |
graph-find-serializers.sh <model> | Serializers for a model |
graph-find-associations.sh <model> | Model associations (belongsto, hasmany) |
graph-find-path.sh <from> <to> | Call path between two functions |
graph-index-delta.sh [project-path] | Full index on first run, then only changed files |
According to the skill, a full index of 1,000+ files takes roughly 30 to 120 seconds, and a delta index of a handful of changed files takes a few seconds.
Quick Start
curl -fsSL https://raw.githubusercontent.com/23blocks-OS/ai-maestro/main/scripts/remote-install.sh | sh
graph-index-delta.sh /path/to/project
graph-find-callers.sh processPayment
Use Cases
- Changing a model and finding every serializer and association that depends on it.
- Changing a function signature and listing every caller first.
- Getting oriented in an unfamiliar codebase.
Limitations
- The component types and association commands reflect Rails-style concepts (models, serializers,
belongs_to), so value varies by stack. - Requires AI Maestro; it is not a standalone code-intelligence server.
FAQ
Is there a separate full-index command?
No. graph-index-delta.sh runs a full index the first time and incremental updates after that.
Alternatives
- Serena provides symbol-level retrieval and references through MCP.
- sentrux, an architectural sensor that reports code structure metrics to coding agents.
Related
Other AI Maestro skills: Memory Search, Docs Search, and Agent Messaging. More in Skills.
Comments
No comments yet. Be the first to comment!
Related Tools
Related Insights
Anthropic Subagent: The Multi-Agent Architecture Revolution
Deep dive into Anthropic multi-agent architecture design. Learn how Subagents break through context window limitations, achieve 90% performance improvements, and real-world applications in Claude Code.
Skills + Hooks + Plugins: How Anthropic Redefined AI Coding Tool Extensibility
An in-depth analysis of Claude Code's trinity architecture of Skills, Hooks, and Plugins. Explore why this design is more advanced than GitHub Copilot and Cursor, and how it redefines AI coding tool extensibility through open standards.
Claude Code's Next Frontier: Not Code, But Your Local Obsidian Knowledge Base
Explore how Obsidian + Claude Code transforms from a knowledge management tool into your private AI assistant. Complete guide including obsidian-skills, Claudian plugin, Claudesidian template, and best practices for achieving both data privacy and AI capabilities.