The Agent Skills Directory

[SAFE]: The skill defines a structured reasoning framework without any malicious instructions.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied input via $ARGUMENTS or the prompt to initiate the deliberation. While this is a common ingestion point for indirect instructions, the skill does not grant excessive privileges or perform dangerous operations with the data.
Ingestion points: SKILL.md (processes user prompt or $ARGUMENTS as the subject of deliberation).
Boundary markers: None explicitly defined within the skill instructions to separate the input from the deliberation protocol.
Capability inventory: The skill utilizes Read, Grep, Glob, Write, WebSearch, and WebFetch tools to facilitate the analysis.
Sanitization: No explicit sanitization or validation of the input is performed within the skill scripts, relying on the platform's native safety filters.

thinking-council