The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface (Category 8) because it ingests untrusted user text describing improvements or errors and uses it to direct sub-agents in modifying DSL configuration files.
Ingestion points: User input strings following triggers like "프로토타입 개선:" or "워크플로우 에러:" are extracted and provided as instructions to the dsl-architect agent in Phase 0.5.
Boundary markers: The skill does not define explicit delimiters or "ignore embedded instructions" warnings when interpolating user input into the prompts for sub-agents.
Capability inventory: The skill possesses the capability to write files to the local filesystem (creating .wip.dsl.yaml and versioned .dsl.yaml files) and orchestrate other agents to interact with the Dify platform.
Sanitization: There is no evidence of validation or sanitization for the user-provided descriptions before they are used to generate or modify the Domain Specific Language (DSL) logic.
[COMMAND_EXECUTION]: The skill is designed for high autonomy, specifically instructing the agent in the "MUST NOT" rules to avoid asking the user questions during its operation. This design reduces human oversight during the automated error-correction loops where the agent modifies and re-runs application logic.

prototype