The Agent Skills Directory

[SAFE]: The skill is entirely descriptive and pedagogical. It defines trigger scenarios based on user cognitive states (e.g., questioning consensus) and provides logical steps for reflection.
[NO_CODE]: There are no scripts, binaries, or tool definitions associated with this skill. It functions as a set of instructions for the AI's persona and logic flow.
[PROMPT_INJECTION]: No attempts to override system prompts or bypass safety guidelines were detected. The instructions focus on improving the user's critical thinking rather than altering the agent's behavior.
[DATA_EXFILTRATION]: There are no network-capable commands (curl, wget), file-read operations on sensitive paths, or hardcoded credentials.
[OBFUSCATION]: A scan for Base64, zero-width characters, and homoglyphs returned no suspicious results. The text is clear and readable.
[DYNAMIC_EXECUTION]: No usage of eval, exec, or runtime compilation patterns.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user input regarding their beliefs, it lacks any capability tools (file write, network, shell) that could be exploited by malicious data.

critical-thinking