adcp

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install the @adcp/sdk package via npm or use npx @adcp/sdk@latest. These resources belong to the vendor's own infrastructure for the Ad Context Protocol.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute adcp CLI commands. These commands perform network requests to AdCP agents and registries, manage local configurations, and run protocol compliance tests.
  • [DYNAMIC_EXECUTION]: The skill recommends the use of npx to fetch and execute the latest version of the AdCP SDK at runtime, ensuring the agent uses up-to-date protocol definitions.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied agent URLs, tool names, and JSON payloads which are interpolated into shell commands.
  • Ingestion points: User input for <agent>, [tool], and [payload] placeholders.
  • Boundary markers: Instructions recommend wrapping payloads in single quotes or using @file.json references to prevent shell misinterpretation.
  • Capability inventory: The skill uses the Bash tool for all operations, including network communication and local file access via CLI flags.
  • Sanitization: The skill relies on standard CLI argument parsing provided by the SDK.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 09:17 AM
Security Audit — agent-trust-hub — adcp