hunter
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external job postings which creates a vulnerability to indirect prompt injection. * Ingestion points: Job postings are fetched from user-provided URLs and public job board APIs (Greenhouse, Lever, Ashby). * Boundary markers: None identified; the skill does not use delimiters to separate instructions from untrusted external content. * Capability inventory: The agent can write files to the local filesystem, perform network fetches, and execute shell commands to open web pages. * Sanitization: No sanitization or filtering is applied to the fetched content before it is used to generate resumes and cover letters.
- [DATA_EXFILTRATION]: The skill collects and stores sensitive Personal Identifiable Information (PII) including full resumes and contact details. * Sensitive user data is stored in the .hunter/ directory. * While the skill is intended for local use, the agent's ability to perform network operations creates a risk surface where this data could be exfiltrated if the agent's logic is manipulated by malicious external content.
- [COMMAND_EXECUTION]: The skill uses operating system shell commands to open URLs. * It utilizes open, xdg-open, and start depending on the platform to launch external job postings in the user's browser.
- [PERSISTENCE]: The skill suggests setting up automated recurring execution. * It provides instructions for the user to create a scheduled task or use a /loop command to run a 'morning' routine daily, which would maintain the skill's activity across sessions.
Audit Metadata