The Agent Skills Directory

[COMMAND_EXECUTION]: The secret resolution example in SKILL.md demonstrates the use of subprocess.check_output to call the op (1Password) CLI tool. This is a common pattern for runtime secret retrieval in development environments and is implemented using a list of arguments rather than a shell string, which is a safer practice.
[INDIRECT_PROMPT_INJECTION]: The skill describes mechanisms for ingesting configuration data from external files (.env, YAML, TOML) and environment variables. While this creates a surface for indirect prompt injection if an agent processes untrusted configuration files, the skill explicitly provides patterns for schema validation and safe loading to mitigate these risks.
Ingestion points: SKILL.md (code examples for loading config files and environment variables).
Boundary markers: None provided in the simplified code examples.
Capability inventory: subprocess.check_output (used for 1Password secret resolution in SKILL.md).
Sanitization: The skill promotes the use of jsonschema for validation and yaml.safe_load for parsing.

configuration-management