learning_ai_common_plat/services
saravanakumardb1 7ed4a105b7 feat(mcp-server): A2A batch-3 — ReflectionSynthesisAgent (mindlyst) + KeyboardDiagnosticsAgent (lysnrai) + NLParserEvalAgent (chronomind)
reflection-synthesis-pipeline.ts: mindlyst.reflections.synthesize
  - ReflectionCollectorAgent -> ThemeExtractionAgent -> SynthesisReportAgent
  - Fetches recent reflections, runs reflection-enrichment extraction, surfaces themes by frequency
  - Proposes weekly summary text; max 52 reflections (1 year lookback)

keyboard-diagnostics-pipeline.ts: lysnrai.keyboard.diagnose
  - KeyboardErrorScannerAgent -> DiagnosticsSessionAgent -> BugReportDraftAgent
  - Scans error telemetry filtered to keyboard surface, groups by anonymousInstallId
  - Opens diagnostics session per erroring install; drafts severity-ranked bug reports
  - 48h default window, configurable minErrors threshold

nl-parser-eval-pipeline.ts: chronomind.nlParser.eval
  - PhraseSamplerAgent -> ParseEvalAgent -> RegressionReportAgent
  - 10-phrase canonical test suite + optional custom phrases
  - Submits to extraction-service timer-parse task, validates parsed field presence
  - Returns pass/fail/partial scorecard with regression list

MCP server total: 102 tools
2026-03-05 16:36:19 -08:00
..
extraction-service fix(jobs): implement stub job handlers with actual functionality 2026-03-02 10:19:15 -08:00
mcp-server feat(mcp-server): A2A batch-3 — ReflectionSynthesisAgent (mindlyst) + KeyboardDiagnosticsAgent (lysnrai) + NLParserEvalAgent (chronomind) 2026-03-05 16:36:19 -08:00
monitoring fix(monitoring): update health-check endpoints for consolidated services 2026-02-17 20:53:37 -08:00
platform-service feat(predictive-analytics): complete admin UI for churn prediction and health scoring 2026-03-03 13:48:37 -08:00