answer-probe
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes 'answers' from external AI engines, which are untrusted sources. This content could contain hidden instructions intended to manipulate the agent during the result review phase.
- Ingestion points: Manual pastes of engine answers as described in
SKILL.mdand API responses handled byaieo/probe.py. - Boundary markers: Absent; there are no instructions to use delimiters or 'ignore' directives to isolate external content from the agent's logic.
- Capability inventory: The skill can execute shell commands (
mkdir,cp,python3) and perform file system writes. - Sanitization: Absent; the agent is directed to manually read and evaluate raw engine responses.
- [COMMAND_EXECUTION]: The skill uses shell commands to initialize its workspace and run the analysis script.
- Evidence: Execution of
mkdir -p aieo/results,cp -n ..., andpython3 aieo/probe.py.
Audit Metadata