vuln-research
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches vulnerability data and advisories from well-known services including NVD (nist.gov), GitHub, CISA, and FIRST.org.
- [COMMAND_EXECUTION]: Uses standard development tools and package managers (npm, pip, go, bundle) via bash to identify installed versions and perform reachability analysis on local source code.
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface: [1] Ingestion points: External vulnerability advisories and GitHub search results are fetched via WebFetch and WebSearch tools as described in the workflow (SKILL.md). [2] Boundary markers: Absent; the skill does not explicitly instruct the agent to disregard instructions that might be embedded in the external content. [3] Capability inventory: The skill has access to Bash, Read, Grep, and Glob tools, which could be leveraged if an agent obeys malicious instructions found in a processed advisory. [4] Sanitization: None identified for external data processing.
Audit Metadata