This directory contains example skill definitions for TinyClaw.
Skills are JSON-defined workflows that combine multiple steps (tool calls, LLM prompts, shell commands) into reusable automation scripts.
Analyze a git repository structure and provide insights.
terraphim-tinyclaw skill save examples/skills/analyze-repo.json
terraphim-tinyclaw skill run analyze-repo repo_path=/path/to/repo analysis_type=structureInputs:
repo_path(required): Path to the git repositoryanalysis_type(optional): Type of analysis - structure, dependencies, or complexity
Steps:
- Get recent git commits
- List repository structure
- LLM analysis with context
Research a topic using web search and summarize findings.
terraphim-tinyclaw skill save examples/skills/research-topic.json
terraphim-tinyclaw skill run research-topic topic="Rust programming" num_results=5Inputs:
topic(required): Topic to researchnum_results(optional): Number of search results to analyze (default: 5)
Steps:
- Web search for topic
- LLM analysis of results
- Log completion timestamp
Perform automated code review on a file or directory.
terraphim-tinyclaw skill save examples/skills/code-review.json
terraphim-tinyclaw skill run code-review target_path=./src/main.rs language=rustInputs:
target_path(required): Path to file or directory to reviewlanguage(optional): Programming language (default: auto-detect)
Steps:
- Read target file(s)
- LLM code review with quality assessment
Generate documentation for a codebase.
terraphim-tinyclaw skill save examples/skills/generate-docs.json
terraphim-tinyclaw skill run generate-docs source_path=./src output_path=./docsInputs:
source_path(required): Path to source code directoryoutput_path(optional): Where to save documentation (default: ./docs/generated)
Steps:
- List source directory structure
- LLM generates documentation outline
- Create output directory
- Write README.md
Basic security scan of files for common vulnerabilities.
terraphim-tinyclaw skill save examples/skills/security-scan.json
terraphim-tinyclaw skill run security-scan scan_path=./src severity=mediumInputs:
scan_path(required): Path to scan for security issuesseverity(optional): Minimum severity to report (default: medium)
Steps:
- Find source files
- Search for potential secrets/credentials
- LLM security analysis
{
"name": "skill-name",
"version": "1.0.0",
"description": "What this skill does",
"author": "Your Name",
"inputs": [
{
"name": "input_name",
"description": "What this input is for",
"required": true,
"default": null
}
],
"steps": [
{
"type": "tool|llm|shell",
...
}
]
}{
"type": "tool",
"tool": "tool_name",
"args": {
"param": "value"
}
}Available tools:
shell: Execute shell commandsfilesystem: Read/write/list filesweb_search: Search the webweb_fetch: Fetch web pagesedit: Edit files with search/replace
{
"type": "llm",
"prompt": "Your prompt here with {input_variable}",
"use_context": true
}{
"type": "shell",
"command": "echo {variable}",
"working_dir": "/optional/path"
}Use {variable_name} syntax to substitute input values:
{
"inputs": [
{"name": "project_name", "required": true}
],
"steps": [
{
"type": "shell",
"command": "mkdir {project_name}"
}
]
}# Save a skill
terraphim-tinyclaw skill save path/to/skill.json
# List all skills
terraphim-tinyclaw skill list
# Load and view a skill
terraphim-tinyclaw skill load skill-name
# Run a skill with inputs
terraphim-tinyclaw skill run skill-name key1=value1 key2=value2
# Cancel a running skill
terraphim-tinyclaw skill cancelSkills are stored in:
- Linux/macOS:
~/.config/terraphim/skills/ - Windows:
%APPDATA%\terraphim\skills\
See the TinyClaw documentation for detailed usage instructions.