The Agent Skills Directory

[NO_CODE] (SAFE): The analyzed file is entirely Markdown and YAML. No scripts, binaries, or system commands are present in the skill definition.\n- [EXTERNAL_DOWNLOADS] (SAFE): No remote resources, packages, or scripts are downloaded. All references are relative paths to local markdown files within the skill package (e.g., workflows/ and references/).\n- [DATA_EXFILTRATION] (SAFE): No network or filesystem operations that could lead to data exposure or exfiltration were detected.\n- [PROMPT_INJECTION] (SAFE): No malicious instruction overrides, jailbreak attempts, or safety filter bypasses are present. The routing logic uses standard intent identification based on user input.\n- [INDIRECT_PROMPT_INJECTION] (INFO): The skill processes user responses to navigate workflows. 1. Ingestion point: User response via AskUserQuestion. 2. Boundary markers: Absent. 3. Capability inventory: No code execution, file writing, or network operations. 4. Sanitization: Absent. Given the lack of actionable capabilities, the risk is negligible.

continuous-discovery