testcontainers
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to inspect local Testcontainers surfaces, including language bindings, package versions, and container runtimes, before performing work. This creates an attack surface where malicious instructions embedded in project code or Dockerfiles could influence the agent's behavior.
- Ingestion points: Local project source code, package manifests (package.json, pom.xml), and Docker/Compose configurations (SKILL.md workflow).
- Boundary markers: None mentioned to separate project data from instructions.
- Capability inventory: Reading local files, suggesting code modifications, and executing shell commands for testing (Docker, Bun, Pip, Go).
- Sanitization: The skill does not explicitly instruct the agent to sanitize or escape data read from the local environment.
- [EXTERNAL_DOWNLOADS]: The skill references official documentation and resources from the Testcontainers project and AtomicJar (the maintainers). This includes documentation sites (e.g., node.testcontainers.org, java.testcontainers.org) and the 'atomicjar/testcontainers-cloud-setup-action' GitHub Action. These are trusted and established sources for the Testcontainers ecosystem.
- [COMMAND_EXECUTION]: The instructions involve running standard development and testing commands, such as package installation (bun add, pip install, go get), Docker operations (docker info), and running test suites (Vitest, Jest, JUnit, pytest) to verify integration behavior.
Audit Metadata