The Agent Skills Directory

[DATA_EXPOSURE]: The skill accesses sensitive local files such as .env and configuration files containing database credentials and secrets. This behavior is the primary intended purpose of the skill (validation and security scanning) and the logic is performed locally.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (configuration files) which could potentially contain malicious instructions. This is mitigated by explicit instructions to parse data safely and avoid executing code from configuration files.
Ingestion points: The ConfigurationAnalyzer class in references/detailed-guide.md reads the content of project files matching patterns like **/*.yaml and **/*.env* using Path.read_text().
Boundary markers: The instructions explicitly direct the agent to "Bound file size/depth, refuse symlink escapes, deduplicate discovery and parse YAML safely."
Capability inventory: The skill possesses file system read capabilities and file watching via chokidar. It does not contain network operations or arbitrary shell command execution capabilities.
Sanitization: The guide mandates "safe" YAML parsing and prohibits the execution of JavaScript configuration files (config.js) to prevent code injection.

deployment-validation-config-validate