The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The ConfigurationAnalyzer class includes logic to search for and read sensitive files such as .env files. It specifically scans for patterns like api_key, password, secret, and aws_access. While this is intended for security validation, it causes the agent to ingest raw credentials and secrets into its context.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from a wide variety of configuration files without implementing boundary markers or sanitization, creating a potential attack surface for indirect instructions embedded in data.
Ingestion points: The ConfigurationAnalyzer._find_config_files and RuntimeConfigValidator.loadAndValidate methods read contents from local files matching patterns like **/*.json, **/*.yaml, and **/*.env.
Boundary markers: There are no instructions provided to the agent to treat configuration data as untrusted or to use delimiters to separate data from instructions.
Capability inventory: The skill possesses the ability to read arbitrary files from the project path, watch files for changes, and perform regular expression matching.
Sanitization: No sanitization or filtering of file content is performed before it is processed by the agent.
[METADATA_POISONING]: The description field in the YAML frontmatter is truncated ("implement configurat"), indicating poorly maintained or incomplete metadata.

deployment-validation-config-validate