← Back to Skills Marketplace
87
Downloads
0
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install nm-gauntlet-graph-build
Description
Build or incrementally update the code knowledge graph for a codebase. Uses Tree-sitter for multi-language AST parsing and stores nodes/edges in SQLite
Usage Guidance
Do not run this skill as-is. Before installing or invoking it, verify where ${CLAUDE_PLUGIN_ROOT}/scripts/graph_build.py comes from and inspect that script: the SKILL.md expects an external Python script that is not included. Ensure python3 and any Tree-sitter bindings it needs are installed and trusted. If you plan to use the skill, either (a) install the referenced claude-night-market/gauntlet plugin from the GitHub homepage listed and confirm CLAUDE_PLUGIN_ROOT is set to that trusted installation, or (b) ask the skill author to bundle the script or declare required binaries and env vars in the skill metadata. Be cautious: running an unreviewed Python script can execute arbitrary code and will write .gauntlet/graph.db and .gauntlet/.gitignore into your repository.
Capability Analysis
Type: OpenClaw Skill
Name: nm-gauntlet-graph-build
Version: 1.0.0
The skill bundle provides instructions for an AI agent to build or update a code knowledge graph using Tree-sitter and SQLite. The instructions in SKILL.md are well-defined, focusing on structural codebase analysis and incremental updates via a local Python script (graph_build.py). No evidence of malicious intent, data exfiltration, or prompt injection was found in the provided files.
Capability Assessment
Purpose & Capability
Name and description match the claimed capability (graph build with Tree-sitter and SQLite). However, the skill requests no binaries, no environment variables, and includes no code. In practice it requires python3, the gauntlet scripts, and Tree-sitter-based tooling; those requirements are not declared, which is disproportionate and inconsistent.
Instruction Scope
SKILL.md instructs the agent to run: python3 ${CLAUDE_PLUGIN_ROOT}/scripts/graph_build.py <dir> (and an incremental mode). That references an undeclared environment variable (CLAUDE_PLUGIN_ROOT) and a script that is not bundled or installed by this skill. The instructions therefore assume external files and env configuration; they also read/write repository files (.gauntlet/graph.db and .gauntlet/.gitignore), which is expected for this purpose but should be explicit.
Install Mechanism
This is an instruction-only skill with no install spec (low installer risk). The problem is omission rather than a risky installer: required scripts are expected to exist elsewhere but the skill does not provide instructions for obtaining or verifying them.
Credentials
No credentials are requested (good), but the instructions reference CLAUDE_PLUGIN_ROOT — an undeclared env var — and implicitly require python3 and likely Tree-sitter bindings. The skill should declare required env vars and binaries; referencing undeclared env state is a coherence and safety concern.
Persistence & Privilege
The skill does not request always:true, does not request persistent/system-wide changes beyond creating .gauntlet files in the repo, and does not alter other skills' configs. Its file writes (graph.db, .gitignore) are consistent with its purpose.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install nm-gauntlet-graph-build - After installation, invoke the skill by name or use
/nm-gauntlet-graph-build - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the graph-build skill, enabling code knowledge graph construction and incremental updates using Tree-sitter for multi-language AST parsing.
- Stores graph data (nodes and edges) in an SQLite database at `.gauntlet/graph.db` with WAL mode for performance.
- Supports full builds and incremental updates based on SHA-256 file hashing to skip unchanged files.
- Extracts rich code relationships (CALLS, IMPORTS_FROM, etc.) from over 20 programming languages.
- Includes user guidance for setup, usage scenarios, and recommended next steps after building the graph.
Metadata
Frequently Asked Questions
What is Nm Gauntlet Graph Build?
Build or incrementally update the code knowledge graph for a codebase. Uses Tree-sitter for multi-language AST parsing and stores nodes/edges in SQLite. It is an AI Agent Skill for Claude Code / OpenClaw, with 87 downloads so far.
How do I install Nm Gauntlet Graph Build?
Run "/install nm-gauntlet-graph-build" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Nm Gauntlet Graph Build free?
Yes, Nm Gauntlet Graph Build is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Nm Gauntlet Graph Build support?
Nm Gauntlet Graph Build is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Nm Gauntlet Graph Build?
It is built and maintained by athola (@athola); the current version is v1.0.0.
More Skills