proffesor-for-testing__agentic-qe/docs/releases/v3.6.10.md

v3.6.10 Release Notes

Release Date: 2026-02-18

Highlights

Completes the QCSD lifecycle with the Production Telemetry Swarm, adds eval-driven development tooling, and integrates Knowledge Graphs into test generation for smarter, faster test creation.

Added

• QCSD Production Telemetry Swarm (#271) — 5th and final QCSD lifecycle phase collecting DORA metrics (deployment frequency, change failure rate, lead time, MTTR) via GitHub API. Includes auto-trigger workflow after npm publish and cross-phase memory hooks (Loop 5: CI/CD to Production).
• Eval-driven development workflow — scripts/eval-driven-workflow.ts for creating, running, and iterating on skill evaluations as the primary development loop.
• Deterministic skill quality scorer — scripts/score-skill-quality.ts computes reproducible quality scores for skills based on structure, coverage, and eval results.
• Skill conflict detector — scripts/detect-skill-conflicts.ts identifies overlapping skill scopes and trigger conflicts across the fleet.
• Eval runner P1 grading improvements — Negative control test cases (inverted pass logic), finding count enforcement, schema validation, weighted grading rubric (completeness/accuracy/actionability), and adaptive rubric with dynamic keyword extraction from prompts.
• KG-assisted test generation — Knowledge Graph integration into all test generators (Jest/Vitest, Mocha, Pytest) with mock declarations emitted from KG dependency analysis. Includes 276x faster HNSW vector loading.

Fixed

• Self-learning hooks not recording from CLI — Hook interactions from CLI commands were silently dropped. All CLI hook interactions now record learning data correctly.

Security

• Dependency updates (#272) — Bumped tar 7.5.7 to 7.5.9 (symlink path traversal fix) and ajv 8.17.1 to 8.18.0 (CVE-2025-69873 ReDoS mitigation).

Upgrade

npx agentic-qe@3.6.10 init --auto