The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it fetches and processes job descriptions from untrusted external URLs to perform keyword optimization and ATS scoring.
Ingestion points: External job description content is retrieved using the WebFetch tool as part of the 'Standalone Quick Optimization' workflow.
Boundary markers: The generation protocol defined in references/resume-protocol.md lacks instructions for using delimiters or boundary markers (e.g., XML tags or triple quotes) when the agent processes external text.
Capability inventory: The skill possesses Write and Edit permissions, which could be exploited by an injection to create or modify local files with malicious content.
Sanitization: There is no evidence of input validation, filtering, or sanitization of the fetched job description data before it is incorporated into the prompt context for keyword extraction.
[EXTERNAL_DOWNLOADS]: The skill documentation references a non-trusted third-party GitHub repository (github.com/erichowens/cv-creator) and describes a deployment process involving external package installation (npm install). This repository is not associated with the skill's author ('curiositech') or the list of trusted vendors.

cv-creator