The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted codebases which may contain malicious instructions embedded in comments or data structures designed to influence the agent's output.\n
Ingestion points: Codebase files located within the user-provided <service-root>.\n
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore embedded instructions in the analyzed files.\n
Capability inventory: The skill executes CLI tools (ripwire, asyncapi), reads multiple file types (source code, YAML, JSON, .env), and writes a generated specification file.\n
Sanitization: There is no explicit sanitization of extracted content before it is included in the generated AsyncAPI YAML.\n- [COMMAND_EXECUTION]: The skill executes shell commands using the ripwire and asyncapi utilities, passing them user-provided directory and file paths.\n
Evidence: Step 1 and Step 7 in SKILL.md involve running ripwire <service-root> and asyncapi validate <file> respectively.\n- [EXTERNAL_DOWNLOADS]: The skill recommends using npx @asyncapi/cli for validation, which involves fetching the package from the public npm registry.\n
Evidence: SKILL.md suggests validating the generated spec using npx @asyncapi/cli if the local CLI is unavailable. This targets a well-known service (npm) and a recognized technology organization (AsyncAPI).

asyncapi-generator