eu-proposal-organisation-assessment
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is composed of process guidelines and markdown templates. It contains no scripts, binary executables, or automated system commands.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze untrusted proposal documents. It mitigates potential injection risks by enforcing a strict evidence mapping process, requiring the agent to cite specific task IDs and page numbers, and instructing the agent to paraphrase rather than directly reproduce bureaucratic prose. The skill does not request network access, file system modifications, or shell execution, limiting the impact of any potential injection.
- [COMMAND_EXECUTION]: No shell commands, system calls, or persistence mechanisms were identified in the instruction files or references.
Audit Metadata