The Agent Skills Directory

[SAFE]: The skill is a meta-analysis tool intended to improve the reliability of AI agents by preventing context saturation and execution of oversized tasks. It implements guardrails by defining explicit 'Stop Conditions' for high-risk activities.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes external data (user requests and repository content) to perform its analysis.
Ingestion points: User input and file content accessed via Read, Grep, and Glob tools.
Boundary markers: None explicitly defined for input data.
Capability inventory: Read, Grep, Glob, and Bash tools are listed in allowed-tools.
Sanitization: None. However, the skill's primary function is to serve as a safety filter, and it explicitly instructs the agent not to execute code or modify files without user confirmation, which mitigates the risk of obeying instructions embedded in analyzed data.

preflight