grimmory-self-hosted-library

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The application described in the skill ingests book metadata from external providers like Google Books and Open Library. This represents a vulnerability surface where untrusted data enters the application context without explicit sanitization shown in the documentation. * Ingestion points: MetadataService.enrichMetadata logic in SKILL.md. * Boundary markers: None specified in the provided code snippets. * Capability inventory: Write access to application data volumes and MariaDB database. * Sanitization: Not demonstrated in the provided examples.
  • [COMMAND_EXECUTION]: The skill provides instructional shell commands for managing Docker containers and resolving file system permission issues (e.g., sudo chown). These are intended for manual execution by a system administrator.
  • [EXTERNAL_DOWNLOADS]: Fetches project source code and Docker images from official repositories, including GitHub and Docker Hub, as part of the standard installation and build process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 08:38 PM
Security Audit — agent-trust-hub — grimmory-self-hosted-library