Interpret Artifacts

{
  "points": {
    "sessionDurationMs": 4500000,
    "promptCount": 47,
    "commandFailRate": 0.12,
    "verifyIntensity": 0.35,
    "manualEditRatio": 0.08,
    "firstChangeLatencyMs": 180000,
    "testResults": {
      "passed": 12,
      "failed": 0,
      "skipped": 2,
      "total": 14
    }
  },
  "processDescriptors": [
    "test-driven",
    "AI-delegating",
    "iterative"
  ]
}

{
  "strategies": {
    "decomposition": {
      "tier": "strong",
      "evidence": "Broke the rate limiter into four distinct subtasks: Redis client setup, sliding window algorithm, middleware wiring, and test coverage.",
      "promptExamples": ["First, let me understand the existing middleware chain..."]
    },
    "contextGathering": {
      "tier": "adequate",
      "evidence": "Read the main router file and config but missed the existing rate limit tests.",
      "promptExamples": ["Read the middleware directory and show me how auth middleware is structured"]
    }
  }
}

{
  "dimensions": {
    "contextManagement": {
      "score": 4,
      "reasoning": "Consistently read relevant files before making changes. Used Grep to find related code patterns."
    },
    "toolSelection": {
      "score": 3,
      "reasoning": "Generally appropriate tool choices, but used Read for large files where Grep would have been more efficient."
    }
  },
  "overallScore": 3.7,
  "summary": "Effective AI tool user with strong context management and room to improve tool selection."
}

{
  "moments": [
    {
      "timestamp": "2026-04-07T12:15:00.000Z",
      "offsetMs": 900000,
      "kind": "breakthrough",
      "summary": "Identified the root cause: the scheduler uses local time for DST calculations instead of UTC",
      "relatedEvents": ["event-uuid-1", "event-uuid-2"]
    },
    {
      "timestamp": "2026-04-07T12:30:00.000Z",
      "offsetMs": 1800000,
      "kind": "pivot",
      "summary": "Abandoned moment.js approach in favor of Temporal API after discovering compatibility issues"
    }
  ]
}

{
  "passed": 12,
  "failed": 1,
  "skipped": 2,
  "total": 15,
  "duration_ms": 4500,
  "details": [
    {
      "name": "should rate limit after threshold",
      "status": "passed",
      "duration_ms": 120
    },
    {
      "name": "should handle concurrent requests",
      "status": "failed",
      "error": "Expected 429, received 200",
      "duration_ms": 350
    }
  ]
}

{
  "headline": "Methodical debugger who systematically isolated the DST issue and delivered a clean fix with comprehensive tests",
  "body": "Alice approached the timezone bug by first reading the scheduler's time handling code and identifying three potential failure points. She narrowed down the root cause within 20 minutes by writing targeted test cases for each hypothesis. Her fix replaced local time arithmetic with UTC-based calculations using the Temporal API, and she added 6 new test cases covering edge cases including spring-forward, fall-back, and timezone boundary transitions. Her prompting style was precise and well-scoped, consistently providing file context and specific requirements."
}

{
  "questions": [
    {
      "question": "You chose the Temporal API over moment.js for timezone handling. What trade-offs did you consider, and how would you handle browser support?",
      "context": "Candidate pivoted from moment.js to Temporal API at minute 30",
      "focus_area": "technical_decisions",
      "difficulty": "medium"
    },
    {
      "question": "Your test suite covers DST transitions but not leap seconds. How would you approach testing calendar edge cases at scale?",
      "context": "Candidate wrote 6 test cases focused on DST",
      "focus_area": "testing_strategy",
      "difficulty": "hard"
    }
  ]
}

artifacts = fetch_artifacts(session_id)
available = {a["kind"] for a in artifacts}

if "prompt_analysis_v1" in available:
    analysis = next(a for a in artifacts if a["kind"] == "prompt_analysis_v1")
    # Process prompt analysis
else:
    # Session may not have enough data, or pipeline is still running
    pass