Kaijan

Plugin kaijan from marketplace kaijan-claude in katalon-studio/kaijan-marketplace. Commands read from the repo's manifest — yesterday.

Skills (18)

• app-map

  Explore a caller-supplied application map through bounded reads and a compact graph. Use to understand a mapped application; not to build a new map or execute browser actions.

• change-risk

  Use Risk Finder for consequential uncertainties in supplied issues/requirements or assumptions in frozen code changes. Use before a PR exists or for code review; not proposed tests, verified defects or execution.

• context-handoff

  Retain relevant work-item, change or test-report context already available during an authorized Loki task, and reference compatible saved input in Loki analysis. Use for context handoff or missing-input follow-up; not workspace collection, test execution or external publication.

• related-work

  Find GitHub issues and pull requests related to an explicit issue, PR, or supplied work item using Context Finder. Use for related-work discovery, not change-risk analysis or test execution.

• test-evidence

  Assess caller-supplied test outcomes and observations against original work-item requirements, with an optional proposed plan. Use after authorized agent execution or for supplied evidence review; not test execution, artifact retrieval or automatic acceptance.

• test-execution

  Execute a complete caller-authorized browser scenario through standalone Runner, read original outcomes and evidence, or Stop that execution. Use only with explicit target/action authority; not for proposing tests or non-browser verification.

• test-planning

  Propose baseline acceptance and risk-focused tests for a supplied issue or work item, optionally with attributed context or findings. Use before a PR exists or for source-backed planning; not test execution or an observed pass.

• assess-change-risk

  Review nontrivial local edits, staged changes, PRs or requirements for consequential assumptions. Use when taking over code or reviewing changes before delivery; skip trivial self-contained edits. Not test generation, verified defects or execution.

• assess-test-evidence

  Assess caller-supplied test outcomes and observations against original work-item requirements, with an optional proposed plan. Use after authorized agent execution or for supplied evidence review; not test execution, artifact retrieval or automatic acceptance.

• find-related-work

  Find GitHub issues and pull requests related to an explicit issue, PR, or supplied work item using Context Finder. Use for related-work discovery, not change-risk analysis or test execution.

• application-setup

  Save the selected application source, original work item and optional target during Kaijan Setup.

• plan-tests

  Propose practical tests for the caller's task and available application, source or change context. Use for test planning, including planning with risk review. Planning stays separate from execution.

• report-feedback

  Help the person send private feedback or prepare a public issue about Kaijan itself in this chat. Use when the person asks to give feedback, make a suggestion, or report a Kaijan bug. Does not report defects in the application being tested.

• review-report

  Read a saved Kaijan Test plan or report, inspect a saved observation, or open a saved browser trace. Use the exact retained report and run without repeating tests or assessment.

• run-test

  Run and assess an authorized browser or macOS desktop test, read its result, or stop the same test. Use for saved Plan tests or a complete supplied scenario. Explicit execution-only requests can use Runner directly.

• setup

  Set up Kaijan and save the selected application context. Use for onboarding, setup and connection readiness. Start with a friendly English welcome.

• test-request

  Use Kaijan to test a selected application or review a change and test its behavior. Also handles requests to review requirements or risks and plan tests. Covers the requested plan, run, assessment, authorized fixes and report.

• work-item

  Check a Kaijan Core connection or read an existing Work Item by its Kaijan UUID. Use for refresh, switch, clearing the conversation's target or disconnecting this Kaijan connection.

MCP servers (2)

• kaijan_core

  runs: ./capabilities/runtime/node ./scripts/core-mcp.mjs

• kaijan_local

  runs: ./capabilities/runtime/node ./scripts/local-mcp.mjs

Similar plugins