goose-integration
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of shell commands to interact with the Goose CLI, including session management and running automated recipes which are YAML-based execution plans.
- [PERSISTENCE]: The skill utilizes the Goose CLI's native support for persistent cron-style scheduling, which allows for recurring task execution that survives session restarts.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where external task descriptions and project instructions are processed by a secondary agent environment. 1. Ingestion points: Task descriptions passed to the goose session command and project context symlinked via .goosehints. 2. Boundary markers: No delimiters or ignore instructions are identified in the provided templates. 3. Capability inventory: Goose sessions have access to Model Context Protocol (MCP) servers, network-based model providers, and local shell extensions. 4. Sanitization: No explicit sanitization or filtering of ingested content is documented.
- [EXTERNAL_DOWNLOADS]: The documentation guides users to install the Goose CLI via Homebrew, a well-known and trusted package management service.
Audit Metadata