The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands using swagger-cli, openapi-generator-cli, yq, and jq to validate and analyze API specifications. It uses standard shell quoting for user-provided file paths, which helps mitigate basic command injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (OpenAPI/Swagger files) and exposes it to the agent's instruction context, creating a potential surface for indirect prompt injection.
Ingestion points: The SPEC_FILE path is passed to shell functions in SKILL.md that read and parse the file content.
Boundary markers: The skill does not employ explicit delimiters or "ignore embedded instructions" warnings when presenting extracted spec data (like descriptions or titles) to the agent.
Capability inventory: The skill has the capability to execute shell commands (swagger-cli, openapi-generator-cli) and perform file system reads across all logic in SKILL.md.
Sanitization: There is no evidence of sanitization, escaping, or filtering of the content extracted from the specification files before it is interpolated into the agent's output.

managing-swagger