triqual-plugin

Classifies test failures using Exolar analytics to determine if they are flakes, bugs, or environmental issues. Trigger when user asks "is this a flake or bug?", wants to analyze recent test failures, asks about test reliability or flakiness, needs failure classification for triage, or has a RUN stage failure documented in a run log. Integrates with run logs.

Analyzes repeated test failures and proposes updates to Quoth documentation and project knowledge.md. Trigger when same error type appears across multiple tests, a fix is applied successfully 3+ times, user asks to "document this pattern" or "add to Quoth", run log has LEARN stage that needs documentation, or session is ending with undocumented learnings. Part of the learning loop: failures -> patterns -> better tests.

Generates test code from plans created by test-planner. Reads run log PLAN stage, project knowledge, and seed files to create well-structured tests. Documents WRITE stage in run log. Trigger when test-planner has created a run log with ANALYZE/RESEARCH/PLAN stages, user says "generate tests from plan", or after test-planner completes (next step in the loop).

Autonomous test healing loop agent. Runs tests, analyzes failures, applies fixes, and loops until tests pass or 25 attempts are reached. After attempt 12, performs deeper analysis with expanded Quoth/Exolar searches. Documents every iteration in the run log. Trigger when test-generator has created test files, tests are failing and need fixing, user says "heal tests" or "fix failing tests", or after test-generator completes (next step in the loop).

Explores applications and creates comprehensive test plans by analyzing requirements, searching Quoth/Exolar, and documenting ANALYZE/RESEARCH/PLAN stages in run logs. Trigger when user asks to "plan tests for {feature}", provides a Linear ticket to test, describes a feature to test, or starts /test command (first agent in the loop).