The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands (git remote get-url origin and git branch --show-current) to determine the repository context. This is a standard and expected method for identifying the project during development workflows.
[EXTERNAL_DOWNLOADS]: The skill fetches build logs and test metadata from CircleCI via the circleci-mcp-server. These interactions are restricted to official CircleCI API endpoints as managed by the MCP server and are necessary for the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of failure logs and test results which could potentially contain malicious instructions designed to influence the agent.
Ingestion points: Data enters the context via mcp__circleci-mcp-server__get_build_failure_logs and mcp__circleci-mcp-server__get_job_test_results in Step 3 and Step 5.
Boundary markers: The instructions do not define specific delimiters for the external log content.
Capability inventory: The agent can read local files, execute git commands, and trigger CircleCI workflow reruns via mcp__circleci-mcp-server__rerun_workflow.
Sanitization: The skill explicitly mitigates risk by instructing the agent to 'Analyze the tool output directly in your response. Do not write scripts or create temporary files to process the results', preventing the execution of code potentially embedded in logs.

debug-ci-failures