The Agent Skills Directory

[SAFE]: No malicious patterns, such as data exfiltration, obfuscation, or unauthorized persistence, were found in the skill.
[COMMAND_EXECUTION]: The skill performs project structure discovery using the Bash tool to run the ls command. This is a legitimate use of the tool to gather context for the documentation scanner.
[PROMPT_INJECTION]: The skill processes project-specific data which introduces a surface for indirect prompt injection. This is mitigated by a human-in-the-loop requirement where the user must approve all changes before the writer subagent acts. * Ingestion points: Project documentation and source code files (SKILL.md Step 1 and 3) * Boundary markers: Absent from subagent briefings * Capability inventory: Includes Bash for discovery and Agent for task delegation * Sanitization: File content is passed to subagents without filtering

update-docs