The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external vendor API documentation and schema files to generate connector definitions.
Ingestion points: The agent ingests element.json, standard-resources/*.json, and external documentation provided by the user to scaffold activities and authentication logic as described in SKILL.md.
Boundary markers: The skill establishes mandatory checks such as 'Inspect before editing' and 'Validate before you finish' to ensure that instructions derived from data are checked against the system's schema requirements.
Capability inventory: The skill leverages Bash to execute the uip CLI tool and uses Write and Edit tools to manage local files. These capabilities are scoped to the development of connectors.
Sanitization: The skill prioritizes using high-level CLI verbs (e.g., activity create, auth set) which enforce structure and validation, rather than raw manual edits to JSON files.
[COMMAND_EXECUTION]: The skill utilizes the uip CLI, a vendor-specific tool for UiPath, to perform connector lifecycle tasks such as initialization, validation, and publishing. The use of this tool is consistent with the skill's primary purpose and author context.
[CREDENTIALS_SAFE]: The instructions include 'Critical Rule 9' which explicitly prohibits the logging, echoing, or hard-coding of secrets. It directs the agent to use the encrypt: true flag for sensitive configuration keys, ensuring credentials are handled securely within the vendor's platform.

uipath-connector-builder