agentic-tests

Plugin agentic-tests from marketplace agentic-tests in yigitkonur/plugin-agentic-tests. Commands read from the repo's manifest — 18 days ago.

Skills (7)

• run

  Run an agentic end-to-end test campaign on web, HTTP API, MCP or mobile targets with fresh Codex subagents per role, blind evidence review and an HTML report. Only when the user invokes it.

• setup

  Prepare a project for agentic-tests once - detect its domains, runtime and runners, and write .claude/agentic-tests.local.md for the agentic-tests run skill. Only when the user invokes it.

• run-agentic-tests

  Use if orchestrating multi-agent E2E campaigns with independent evidence review, isolated runtimes, and defect fix/retest loops.

• ego-browser

  Default browser for any web task: open sites, sign in, fill forms, screenshot, extract data, test or QA web apps. Prefer over other web tools.

• test-by-maestro

  Use if writing, running, or debugging Maestro E2E test flows on iOS Simulators, Android, or Web, including MCP agent automation and CI suites.

• test-by-mcpc-cli

  Use skill if you are driving mcpc 0.7.x to test or smoke-check an MCP server over stdio or Streamable HTTP.

• run-forked

  Run an agentic-tests campaign in a forked Sonnet orchestrator. Your session waits and receives only the final report.

Subagents (8)

• diagnostician

  Explains one verified test failure, separates product defects from test or environment problems, and proves or rules out a root cause in the source. Dispatched only by the agentic-tests orchestrator with a handoff file.

• implementer

  Fixes one confirmed defect from an agentic-tests campaign in an isolated git worktree, runs developer checks, and opens a pull request. Dispatched only by the agentic-tests orchestrator with a handoff file.

• integrator

  Serially merges one implemented fix from an agentic-tests campaign into the target branch and records the new target and retest obligations. Dispatched only by the agentic-tests orchestrator with a handoff file.

• orchestrator

  Runs a whole agentic-tests campaign in the background for /agentic-tests:run --detach, dispatching the expert agents and reporting once at the end. Launched only by the agentic-tests run skill.

• plan-auditor

  Independently audits an agentic-tests plan before execution, or the campaign's recorded proof at closure, from files only. Dispatched only by the agentic-tests orchestrator with a handoff file.

• planner

  Builds the coverage, dependency and resource plan for an agentic-tests campaign's frozen cases, for an independent audit. Dispatched only by the agentic-tests orchestrator with a handoff file.

• ticket-writer

  Publishes one deduplicated GitHub issue for a confirmed, in-scope product defect found by an agentic-tests campaign. Dispatched only by the agentic-tests orchestrator with a handoff file.

• verifier

  Blind evidence reviewer for one executed test case. Opens saved screenshots, logs and JSON, judges each frozen expectation, never touches the app. Dispatched only by the agentic-tests orchestrator with a handoff file and slot.

Similar plugins