agent-communication

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION] (SAFE): The instructions in SKILL.md define a specific persona and prioritize provided reference files. While it instructs the agent to 'Ignore generic approaches,' this is context-specific to the domain of agent communication and does not attempt to bypass global safety filters or extract system prompts.
  • [DATA_EXFILTRATION] (SAFE): No hardcoded credentials, sensitive file paths, or network-enabled commands (e.g., curl, wget) are present in the files.
  • [REMOTE_CODE_EXECUTION] (SAFE): The skill does not define external dependencies or attempt to download/execute remote scripts.
  • [INDIRECT_PROMPT_INJECTION] (LOW):
  • Ingestion points: The skill is designed to handle 'agent communication' and 'message passing' (SKILL.md).
  • Boundary markers: No specific boundary markers for untrusted agent data are defined.
  • Capability inventory: No active tools or subprocess capabilities are present in the skill files.
  • Sanitization: The skill explicitly mitigates injection risks by requiring 'schema validation' (references/validations.md) and 'constrained, validated communication' (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:03 PM