astrofy-setup
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run local CLI commands, specifically
astrofy setup, to manage project state and coordination. - [INDIRECT_PROMPT_INJECTION]: The skill reads external project files like
package.json,astro.config.*, and local markdown instructions, which are then used to determine the workflow and parameters for CLI commands. This constitutes an indirect prompt injection surface. - Ingestion points: Configuration files (
package.json,astro.config.*), Git state, and local project documentation. - Boundary markers: No delimiters or "ignore" instructions are provided for ingested content in the instructions.
- Capability inventory: Execution of
astrofyCLI commands and project file modification. - Sanitization: No explicit validation or filtering of ingested file content is described.
- [CREDENTIALS_UNSAFE]: The project workflow defined in
workflow.jsonincludes a security audit step (astrofy-security) that processes sensitive data such as application secrets and session information. While appropriate for a security-focused tool, it involves the handling of potentially sensitive project credentials.
Audit Metadata