call-adcp-agent

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill exclusively references official vendor domains (adcontextprotocol.org) and repositories (github.com/adcontextprotocol) for documentation and protocol schemas.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external tool responses to manage protocol state and error recovery.
  • Ingestion points: structuredContent (MCP) and task.artifacts (A2A) tool outputs from AdCP agents.
  • Boundary markers: The instructions do not explicitly define boundary markers for the data, relying on structured JSON shapes.
  • Capability inventory: The skill allows the agent to perform mutating operations like create_media_buy, sync_creatives, and activate_signal.
  • Sanitization: No specific sanitization of the tool output content is described; the agent is expected to parse the JSON error pointers directly.
  • [EXTERNAL_DOWNLOADS]: Mentions downloading protocol schemas via a tarball from the official vendor site. This is a standard configuration step for SDK synchronization and represents normal vendor functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 10:47 AM