docker-expert
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various shell commands including
docker,docker-compose, andfindto analyze project structures and validate the status of containers (SKILL.md). \n- [REMOTE_CODE_EXECUTION]: The skill facilitates the execution of build instructions and runtime logic throughdocker buildanddocker runcommands when validating local Dockerfiles and images (SKILL.md). \n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes contents from user-controlled configuration files without isolation. \n - Ingestion points: The skill searches for and reads
Dockerfile*and*compose*.ymlfiles in the local workspace (SKILL.md). \n - Boundary markers: None. The skill does not implement delimiters or instructions to ignore embedded commands within ingested files. \n
- Capability inventory: The skill has access to high-capability tools such as
docker build,docker run, anddocker exec(SKILL.md). \n - Sanitization: There is no evidence of content filtering or validation for the configuration files it processes. \n- [EXTERNAL_DOWNLOADS]: The skill interacts with external services to pull container images and run security audits using Docker Scout, representing standard interactions with well-known container ecosystem services (SKILL.md).
Audit Metadata