What problem does it solve? Robot services (Issue bots, PR bots, auto-assign/label bots) have no UI and only react to webhook events from code hosting platforms, so their test cases cannot be designed without a real repository and real platform behavior. This Skill enforces a probe-first workflow that calls the platform API before writing any case, preventing test expectations based on outdated or wrong design-document assumptions. ## Core Features & Use Cases - Mandatory repository gate: Refuses to generate cases until the user provides an execution repository URL and token, since robot behavior can only be verified against real webhook events. - Platform API probing: Detects repo reachability, members, bot account identity, default assignee, issue_type dictionaries, comment language, real HTTP status codes, and dashboard capabilities across gitcode, gitee, github, atomgit, and gitlab. - Dual-use case output: Produces 8-field test case tables (importable to ZenTao/Tapd/Jira) each paired with an agent-exec YAML block containing curl steps, wait actions for async webhook processing, and jsonpath assertions filtered by bot account. - Use Case: Given a gitcode repository URL and token, generate a full regression suite for an auto-assign robot covering default assignment, /assign and /unassign comment commands, permission rejections, and auto-labeling, with all assertions grounded in probed platform facts. ## Quick Start Generate robot test cases for the auto-assign bot using repository https://gitcode.com/openeuler-test/test-feature with my access token.
