internal-comms

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection by encouraging the agent to process untrusted data from Slack, Google Drive, and Email without safety boundaries.
  • Ingestion points: Untrusted data enters via company communication tools as described in the example guideline files (3p-updates.md, company-newsletter.md, faq-answers.md).
  • Boundary markers: No delimiters or instructions to ignore embedded commands are provided to separate fetched content from the system instructions.
  • Capability inventory: The skill contains no code or scripts but utilizes the agent's inherent ability to read, summarize, and format external data for communication.
  • Sanitization: No content validation, escaping, or sanitization is specified for the data retrieved from external tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 01:13 AM