This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+14 more
Summary
AI summaryUpdates MCP Tasks extension, Release pipeline, and Housekeeping across a mixed release.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
Adds experimental MCP Tasks protocol for asynchronous long-running runs. Adds experimental MCP Tasks protocol for asynchronous long-running runs. Source: llm_adapter@2026-07-12 Confidence: high |
— |
| Feature | Low |
Propagates W3C trace-context to spawned agent per task. Propagates W3C trace-context to spawned agent per task. Source: llm_adapter@2026-07-12 Confidence: high |
— |
| Bugfix | Medium |
Makes MCP registry listing publish idempotent, treating already‑live versions as success. Makes MCP registry listing publish idempotent, treating already‑live versions as success. Source: llm_adapter@2026-07-12 Confidence: high |
— |
Full changelog
v3.4.4
Released 2026-07-12.
A capability release that adds the experimental MCP Tasks protocol surface to
the Bernstein MCP server, plus release-pipeline hardening.
MCP Tasks extension
-
Long-running runs are now exposed through the experimental MCP Tasks protocol.
A task-capable MCP client (Claude Code, Cursor, Cline, and others) can start a
run withbernstein_run, receive aCreateTaskResult, and then poll status
and retrieve the result asynchronously through theget_task,
get_task_result,list_tasks, andcancel_taskhandlers without holding a
blocking session open. The task handle embeds the run's audit-chain head hash
({task_id}:{head_hash}) so a stateless client can still tie a progress claim
back to the signed chain. -
W3C trace-context (
traceparent,tracestate,baggage) is ingested from
the incoming request and propagated to the spawned agent, scoped per task so a
run without trace-context never inherits another run's context. -
Every task status projects to a terminal MCP status on completion, and
get_task_resultreports an error for terminal-error and in-flight tasks
instead of reporting success.Thanks to @Amanmeena0 for contributing the MCP Tasks protocol implementation
and the trace-context propagation.
Release pipeline
- The MCP registry listing publish is now idempotent: a version that is already
live is treated as success instead of failing the release on a re-run or a
tag-push/dispatch race. (#2466)
Housekeeping
- Refreshed the packaged adapter last-green projection and its docs table from
the nightly conformance canary; every row stays anchored to its attesting
receipt.
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
About chernistry/bernstein
Deterministic multi-agent orchestrator for 18 CLI coding agents (Claude Code, Codex, Cursor, Aider, Gemini CLI, OpenAI Agents SDK, and more). MCP server mode (stdio + HTTP/SSE) exposes the orchestrator to any MCP client. Git worktree isolation per agent, HMAC-chained audit trail, cost-aware model routing via contextual bandit. ~11K monthly PyPI downloads, Apache 2.0.
Related context
Related tools
Earlier breaking changes
- v3.7.1 `bernstein approve` and `bernstein reject` now enforce identifier regex `[A-Za-z0-9._-]{1,64}`.
- v3.7.1 Tampered mission ledger reports as unverified rather than not-found.
- v3.7.1 `mission define` now refuses phases without gate tasks.
- v3.5.0 MCP client, transport, and gateway become stateless; calls carry content‑derived trace IDs in _meta.
Beta — feedback welcome: [email protected]