The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the configuration of https://rube.app/mcp as an MCP server. This third-party endpoint provides the necessary tool definitions and logic for CircleCI integration.
[COMMAND_EXECUTION]: Includes tools for triggering remote pipelines (CIRCLECI_TRIGGER_PIPELINE) and managing authentication through RUBE_MANAGE_CONNECTIONS.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the processing of external build data. Ingestion points: Retrieves artifacts and test metadata via CIRCLECI_GET_JOB_ARTIFACTS and CIRCLECI_GET_TEST_METADATA. Boundary markers: No instructions are provided to delimit or sanitize external data from agent commands. Capability inventory: The agent can trigger pipelines and manage tool connections based on its interpretation of the ingested data. Sanitization: No sanitization or validation of external content is described.

circleci-automation