The Agent Skills Directory

[COMMAND_EXECUTION]: The shell script scripts/validate_all_skills.sh executes a Python validator located at a path determined by the SKILL_VALIDATOR environment variable or a default path in the user's home directory. This allows for the dynamic execution of scripts from computed paths.
[PROMPT_INJECTION]: The skill acts as a management layer that ingests and processes other SKILL.md files, which constitutes a surface for indirect prompt injection.
Ingestion points: scripts/validate_all_skills.sh recursively finds and processes skill files within a specified repository root.
Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the script's processing logic.
Capability inventory: The skill executes filesystem operations and shell commands through its validation script.
Sanitization: The skill does not implement explicit sanitization of the content found within the processed SKILL.md files before passing them to the validator.
[SAFE]: The skill's documentation includes a security review section that provides best practices for auditing skills, such as checking for hardcoded credentials and blocking dangerous command patterns.
[SAFE]: No hardcoded secrets, remote code downloads from untrusted sources, or data exfiltration patterns were detected in the skill's own implementation.

ai-skill-manager