to-spec
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill synthesizes technical specifications from conversation history, which is a potential surface for indirect prompt injection.
- Ingestion points: The skill is instructed in
SKILL.mdto take the "current conversation context" as input. - Boundary markers: Absent; there are no instructions to use delimiters or to ignore potential instructions embedded within the processed conversation data.
- Capability inventory: The agent is authorized to perform repository exploration and publish to an external issue tracker.
- Sanitization: Absent; the skill does not define filtering or escaping logic for the conversation content.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill is designed to transmit data to an external system.
- Evidence:
SKILL.mdexplicitly directs the agent to "publish it to the project issue tracker." While this is the intended primary purpose, it involves moving internal conversation context to an external project management tool.
Audit Metadata