convening-experts
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWNO_CODE
Full Analysis
- NO_CODE (SAFE): The skill consists entirely of Markdown-based instructions and reference files. No scripts, binaries, or automated tasks are included within the skill package.
- COMMAND_EXECUTION (SAFE): No shell commands, subprocess calls, or system-level execution patterns were identified in the analyzed content.
- DATA_EXFILTRATION (SAFE): No network requests (e.g., curl, wget) or external communication methods are defined. All references are local to the skill's filesystem.
- PROMPT_INJECTION (SAFE): The instructions provide a structured reasoning framework for the agent to follow. There are no attempts to override system prompts, bypass safety filters, or ignore previous instructions.
- Indirect Prompt Injection (SAFE): The skill processes user queries by adopting specific expert personas. Because the skill lacks any capabilities to write files, execute commands, or access the network, the impact of potential malicious content within user input is limited to affecting the generated text output (display only).
Audit Metadata