enrich

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill's protocol explicitly instructs the agent to send existing internal context ("brain knowledge") to external third-party services (such as Perplexity, Brave Search, and Exa) to identify updates. This pattern involves exposing internal data to external search providers.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes information from untrusted external sources, including web research, social media posts, and public APIs, and writes this data into the internal knowledge base.
  • Ingestion points: Data is retrieved from various external sources detailed in Step 4 (Web research, Social media lookup, Enrichment APIs).
  • Boundary markers: Absent. The instructions do not specify the use of delimiters or markers to isolate untrusted external content from the agent's instructions.
  • Capability inventory: The skill utilizes put_page, put_raw_data, and add_timeline_entry to write processed external data into the file system/brain.
  • Sanitization: Absent. There are no instructions to sanitize, filter, or validate the content retrieved from external sources for malicious prompts before storage.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 06:18 PM
Security Audit — agent-trust-hub — enrich