agents-platform
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources, which is a common attack surface for indirect prompt injection.
- Ingestion points: The workflow involves fetching external web content via
agents_create_kb_urland reading potentially user-influenced conversation transcripts throughagents_get_conversation(SKILL.md). - Boundary markers: The skill includes operational guardrails, such as instructing the agent to confirm deletions with the user and use only verified IDs, but it does not specify explicit delimiters or "ignore" instructions for content retrieved from external URLs (SKILL.md).
- Capability inventory: The skill possesses broad capabilities to modify account states, including creating/updating agents, managing knowledge base content, and triggering deployments via the
agents_*toolset (SKILL.md). - Sanitization: The instructions do not define specific sanitization, filtering, or validation steps for content retrieved from knowledge base sources or conversation history before it is processed by the agent (SKILL.md).
Audit Metadata