Skip to content

claude-flow

v3.10.8 Bugfix

This release fixes issues for SREs watching stability and regressions.

Published 5d AI Agents & Assistants
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agentic-ai agentic-framework agentic-rag agentic-workflow agents ai-agents
+14 more
ai-assistant ai-coding ai-skills autonomous-agents claude-code codex mcp-server multi-agent multi-agent-systems npm skills swarm swarm-intelligence typescript

Summary

AI summary

Immediate cache invalidation on Q‑LearningRouter updates and correct parsing of --explore false flag.

Changes in this release

Bugfix Medium

Invalidates route cache entry immediately after QLearningRouter.update().

Invalidates route cache entry immediately after QLearningRouter.update().

Source: llm_adapter@2026-05-29

Confidence: high

 —
Bugfix Medium

Correctly parses `--explore false` to disable exploration flag.

Correctly parses `--explore false` to disable exploration flag.

Source: llm_adapter@2026-05-29

Confidence: high

 —
Full changelog

Ruflo v3.10.8 — routing-learning correctness fixes

Two follow-up fixes from the intelligence audit's remaining punch-list (docs/reviews/intelligence-system-audit-2026-05-29.md).

Bug B — stale route cache hid learning

QLearningRouter.update() only invalidated the whole route cache every 50 updates, so a freshly-learned Q-update stayed hidden behind a stale cached decision — feedback appeared to have no effect on routing in-process until 50 updates accumulated. Now the updated state's cache entry is invalidated immediately. Verified: learned route flips coder→researcher within 10 updates.

Bug C — --explore false was ignored

Boolean flags dropped an explicit space-form value, so a default-true boolean (explore) stayed true even with --explore false — exploitation could not be forced. The parser now consumes a true/false literal for boolean flags (--explore false, -e false); --explore=false and --no-explore keep working. Verified: deterministic exploitation with explore=false.

Deferred (documented in the audit, not patched)

  • SONA default-path: re-examined — the default path already learns via LocalSonaCoordinator; the inert piece is a non-load-bearing supplementary forward.
  • WASM MicroLoRA apply() inert — lives in the @ruvector/ruvllm dependency (needs upstream fix).
  • Per-task bandit priors — changes a persisted schema (needs an ADR + migration).

All three packages published at 3.10.8 (latest/alpha/v3alpha in lockstep). CI: 31/31 green.

🤖 Generated with RuFlo

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track claude-flow

Get notified when new releases ship.

Sign up free

About claude-flow

Deploy multi-agent swarms with coordinated workflows.

All releases →

Related context

Beta — feedback welcome: [email protected]