docker-compose-orchestration
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFEPRIVILEGE_ESCALATIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: Multiple documentation examples contain hardcoded default passwords and secrets instead of using environment variables or secret management practices.
SKILL.mduses values likeJWT_SECRET=dev-secret,POSTGRES_PASSWORD=password, andPGADMIN_DEFAULT_PASSWORD=admin.EXAMPLES.mdincludes examples withMONGO_INITDB_ROOT_PASSWORD=secret,ME_CONFIG_MONGODB_ADMINPASSWORD=secret, andMYSQL_ROOT_PASSWORD=root_secret.README.mdfeaturesPOSTGRES_PASSWORD=secretin the primary example configuration.- [PRIVILEGE_ESCALATION]: Configuration templates for several infrastructure tools mount sensitive host paths into containers, creating a potential vector for host compromise.
- Examples for GitLab Runner, Jenkins, Traefik, and Portainer in
SKILL.mdandEXAMPLES.mdmount the host Docker socket (/var/run/docker.sock) to allow the services to manage other containers. - Monitoring tool configurations (Node Exporter, cAdvisor) mount the host root filesystem (
/) to provide host-level metrics. - [EXTERNAL_DOWNLOADS]: The skill involves pulling container images from established public registries.
- Fetches official images including
nginx:alpine,postgres:15-alpine, andredis:7-alpinefrom Docker Hub, a well-known service. - [INDIRECT_PROMPT_INJECTION]: The skill provides numerous infrastructure templates that an agent could use to generate code, which represents an attack surface for indirect injection if data provided by the user is not sanitized.
- Ingestion points: The skill uses YAML configuration templates across all documented files (README.md, SKILL.md, EXAMPLES.md).
- Boundary markers: Not present; the skill lacks specific delimiters or warnings to ignore instructions embedded in user-supplied configuration data.
- Capability inventory: Includes the ability to execute shell commands via
docker composefor building, running, and managing multi-container stacks. - Sanitization: No explicit methods for escaping or validating external content before it is interpolated into the generated Compose files.
Audit Metadata