The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/validate_local_compose.py executes docker compose config via subprocess.run to validate generated files. The command is constructed using a list of arguments, which is a standard method for safe interaction with system binaries.- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted project files such as Compose files and Dockerfiles to generate development configurations, creating a potential surface for indirect instructions.
Ingestion points: Project-level files (Compose, Dockerfile, and environment examples) as specified in the SKILL.md workflow.
Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore embedded content when reading project files.
Capability inventory: The skill can execute the local docker CLI via the provided Python script.
Sanitization: Absent; project content is analyzed for configuration structure but is not sanitized for potential prompt instructions.

configure-local-compose