career-companion

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow ingests and extracts requirements from Zero G Talent job postings (“live openings”) by calling search_jobs//api/agent/jobs?… and then fetching the selected job’s full description via get_job/{jdUrl}?format=md, meaning outsider-authored job/JD text can be processed at runtime.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime curl/MCP calls to Zero G Talent APIs (e.g. https://zerogtalent.com/api/agent/jobs and https://zerogtalent.com/api/agent/job/{slug}) to fetch full job descriptions which are then used to generate prompts (resume tailoring and interview questions), so these external endpoints directly control agent outputs.