The Agent Skills Directory

[SAFE]: The skill consists exclusively of markdown documentation and instructional prompts for an AI agent. No executable scripts (.py, .js, .sh), shell commands, or configuration files were found.
[SAFE]: No remote code execution or external downloads were identified. All references are local to the skill's file structure.
[SAFE]: There are no requests for sensitive file access, environment variable exposure, or network exfiltration operations.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted user content (such as arguments or plans for analysis), it lacks any executable capabilities or tools. Consequently, any malicious instructions embedded in processed data cannot result in system-level compromise or exfiltration within this skill context.
Ingestion points: User-provided text or plans triggered by critical thinking requests.
Boundary markers: None explicitly defined in the prompts.
Capability inventory: No tools, subprocess calls, or network operations are used.
Sanitization: None required as the skill possesses no sensitive capabilities.

critical-thinking