The Agent Skills Directory

[SAFE]: No security issues detected. The skill is designed for codebase archaeology and onboarding, utilizing standard local search tools like ripgrep, find, and cat. All commands are restricted to the repository context and intended for technical analysis.
[COMMAND_EXECUTION]: The skill uses shell commands to analyze the repository structure and code patterns. These commands (e.g., rg, find, cat) are used exclusively for read-only exploration and do not perform any destructive or unauthorized operations.
[PROMPT_INJECTION]: The skill includes instructions for managing subagents and synthesizing their output. It incorporates a validation process (Specialist Packet Contract) to ensure the agent uses verified evidence and avoids incorporating low-quality or misleading information into the final architecture summary.
[CREDENTIALS_UNSAFE]: While the skill searches for configuration files and environment variable usage to understand system boundaries, it explicitly instructs the agent to avoid exposing actual secret values in its reporting.

codebase-archaeology