agent-qa-result-triage skill

Agent QA Result Triage

Overview

Classify a failed Agent QA run from its recorded evidence instead of guessing. Inspect the run, steps, artifacts, and logs; choose one fixed category; and return confidence, likely ownership, and the next evidence-backed action.

When to Use

• Investigating a failed or interrupted Agent QA run.
• Inspecting run artifacts, step results, or execution logs.
• Comparing recent related runs for recurring failure patterns.
• Deciding whether a failure belongs to a test, product, hook, browser/mobile runtime, or infrastructure owner.

Workflow

1. Start with agentqaget_run for run status, suite child context, steps, and attempts.
2. Fetch evidence before deciding:

• agentqagetrunartifact
• agentqagetrunsteps
• agentqagetrunlogs
• agentqagetrunexecution_logs

1. Call agentqaclassify_failure and use its category as the default classification unless stronger evidence contradicts it.
2. Compare recent related runs when they are available in the classifier output.
3. Return a concise triage result: category, confidence, evidence, likely fix area, and next action.
4. For code changes, switch to agent-qa-debug-fix after triage is complete.

Categories

Use exactly one category from references/triage-categories.md:

• timeout
• appium_startup
• browser_disconnect
• elementnotfound
• assertion_failure
• hook_failure
• infrastructure
• unknown_failure

Evidence Rules

• Quote or summarize concrete artifact, log, or step evidence.
• Mention missing artifact sections when they limit confidence.
• Do not invent screenshots, videos, logs, or memory context that MCP did not return.
• If MCP is unavailable, use dashboard REST APIs or Agent QA CLI output as a fallback and state which evidence was unavailable.
• Redact credentials, session tokens, personal data, and unrelated application content from the report.

Example

{
  "category": "element_not_found",
  "confidence": "high",
  "evidence": ["Step 4 could not resolve the described checkout button"],
  "likely_fix_area": "test definition or changed product UI",
  "next_action": "Inspect the captured UI context, then compare the current checkout screen"
}

Limitations

• Classification is only as reliable as the retained run artifacts and logs.
• A failure category identifies the most likely failure surface; it does not prove root cause.
• Missing screenshots, DOM/accessibility context, device logs, or prior runs must lower confidence.
• This skill does not modify tests or application code; use agent-qa-debug-fix for an authorized repair.