The Agent Skills Directory

[SAFE]: A comprehensive audit of the skill instructions and templates revealed no evidence of malicious intent, obfuscation, or hidden logic. The skill adheres to the 'never fabricate' principle for user data.\n- [COMMAND_EXECUTION]: The skill workflow involves saving generated HTML files to the local file system. While it references a specific path on the author's machine (/Users/yanliu/Desktop/Claude skills/), it provides for graceful fallback to the current directory for other users.\n- [EXTERNAL_DOWNLOADS]: In prompts/linkedin-import.md, the skill uses WebFetch to retrieve LinkedIn profile data. This is a standard and necessary function for the skill's purpose of importing professional history.\n- [PROMPT_INJECTION]: The skill processes data from external sources such as LinkedIn profiles and user-uploaded files. This represents an indirect prompt injection surface. Ingestion points: prompts/linkedin-import.md (WebFetch), prompts/beautify.md (User files). Boundary markers: None explicitly defined. Capability inventory: File writing and HTML generation. Sanitization: Relies on agent parsing.\n- [NO_CODE]: The skill contains no executable agent scripts (Python/Node.js). It relies entirely on instruction-based prompts and static HTML templates with minimal client-side JavaScript for basic UI functions.

resume-skill