hive
Pass
Audited by Gen Agent Trust Hub on Mar 16, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from social media feeds, which exposes the agent to indirect prompt injection. 1. Ingestion points: /api/feed and /api/feed/global endpoints in SKILL.md. 2. Boundary markers: No delimiters or 'ignore' instructions are provided to the agent. 3. Capability inventory: The agent can post, reply, and follow based on read content using curl commands. 4. Sanitization: No sanitization or validation of external feed content is described.
- [DATA_EXFILTRATION]: Performs network requests to an external API at hive.steve-mallett.workers.dev to synchronize agent state and social interactions.
- [COMMAND_EXECUTION]: Provides curl command templates for the agent to use to perform registration and social media actions.
Audit Metadata