The Agent Skills Directory

[NO_CODE] (SAFE): The skill consists exclusively of Markdown files providing procedural instructions. No executable scripts (.py, .js, .sh), binary files, or dependency manifests (package.json, requirements.txt) are included.\n- [COMMAND_EXECUTION] (SAFE): There are no patterns suggesting the execution of system commands, shell scripts, or unauthorized subprocess calls.\n- [DATA_EXFILTRATION] (SAFE): No access to sensitive file paths (e.g., ~/.ssh, .env) or unauthorized data transmission to external domains was detected. The use of web search in Phase 3 is a standard, non-malicious tool capability for verifying concepts.\n- [PROMPT_INJECTION] (SAFE): No malicious override markers, bypass instructions, or jailbreak patterns were identified. The surface for indirect prompt injection is inherent to the conversational nature of the skill and does not pose a specific threat.

thought-mining