The Agent Skills Directory

[SAFE]: No malicious patterns, such as prompt injection or unauthorized code execution, were detected. The skill is designed for legitimate architecture management.- [COMMAND_EXECUTION]: The skill uses curl commands to interact with the Revision REST API. These operations are well-documented, standard for API clients, and used for their intended purpose.- [EXTERNAL_DOWNLOADS]: All network interactions target organization-specific subdomains of revision.app. This is a vendor-owned resource corresponding to the skill author 'revision-org'.- [DATA_EXFILTRATION]: The skill handles authentication by instructing the agent to request an API key and organization URL from the user. It does not attempt to access local sensitive files or exfiltrate data to unauthorized third-party services.- [PROMPT_INJECTION]: The skill processes component and diagram data from the Revision API, creating an indirect prompt injection surface. Ingestion points: GET endpoints for components and diagrams in OPENAPI.md. Boundary markers: Absent. Capability inventory: curl POST/PATCH operations for data mutation in SKILL.md. Sanitization: Absent. This is considered standard behavior for an API integration tool.

revision-external-api