The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines tools that execute local Python scripts (scripts/project_scaffolder.py and scripts/code_quality_analyzer.py) to automate development tasks such as project generation and codebase auditing.
[DYNAMIC_EXECUTION]: The Project Scaffolder tool generates project boilerplate, including configuration files and scripts, based on pre-defined templates for various technology stacks like Next.js, MERN, and FastAPI.
[CREDENTIALS_UNSAFE]: The Code Quality Analyzer tool is designed to scan projects for hardcoded secrets and security vulnerabilities. This functionality involves accessing potentially sensitive configuration files (e.g., .env files) to perform its audit.
[INDIRECT_PROMPT_INJECTION]: The skill's ability to analyze external codebases creates a surface for indirect prompt injection, where malicious instructions embedded in the analyzed source code could influence the agent's summary or interpretation of the audit results.
Ingestion points: The project directory path provided to scripts/code_quality_analyzer.py.
Boundary markers: No specific delimiters or instructions to ignore embedded instructions are defined in the skill's usage guides.
Capability inventory: The skill can execute Python scripts, perform file system operations (list, copy), and run package manager commands.
Sanitization: The documentation does not describe explicit sanitization or validation of the content within the codebases being analyzed.

senior-fullstack