internal-comms

Pass

Audited by Gen Agent Trust Hub on Apr 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) due to its design of ingesting and summarizing data from potentially untrusted external sources.
  • Ingestion points: The workflow in examples/3p-updates.md, examples/company-newsletter.md, and examples/faq-answers.md explicitly instructs the agent to gather information from Slack threads, emails, Google Drive documents, and external press materials.
  • Boundary markers: The instructions lack delimiters or specific warnings to ignore embedded commands, which may cause the agent to accidentally follow instructions found within the summarized data.
  • Capability inventory: The skill possesses the ability to read sensitive company communications and generate output for distribution across internal channels (Slack, Email), creating a path for manipulated content to reach a wide audience.
  • Sanitization: No logic is present for sanitizing, escaping, or validating the input data before it is processed into the final communication format.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 14, 2026, 06:35 PM