Repository navigation
Bernstein 4.0.0
At a glance
- If a workflow node fails, Bernstein skips later layers, including independent branches, as in
bernstein workflow resume. A detached SSH task with a nonzero remote exit fails the run and leaves later tasks pending. bernstein govern audit --strictgives CI distinct exit codes for failed checks and unmeasurable prerequisites.- At run close, Bernstein seals the final delegation digest, hop count and manifest configuration with the install audit key.
- Fleet selection identifies an enabled instance where one worker matches the task type, dispatch tier and required capabilities.
What is partial in 4.0.0 summarizes selected features and their limits.
Before you upgrade
Read the 3.24 to 4.0 migration guide.
- Quality gates. Previously ignored settings such as
quality_gates.large_file_checkinbernstein.yamlnow apply. A value of the wrong type raisesSeedError(#6615). - Delegation verification. If a manifest declares a final head with a missing or invalid seal,
bernstein delegation verifyexits1. This includes heads declared before seals existed. An otherwise intact, nonempty chain with no final declaration reportsvalid=falseand exits3. An interrupted close can leave the declaration missing (#5352). - MCP versions. Bernstein uses SDK 2.3.x with the 2025-11-25 task wire contract. By default, the HTTP transport returns 400 for an unsupported
MCP-Protocol-Versionheader. During the handshake, the server answers an unknowninitialize.params.protocolVersionwith its default revision. - Quarantine. The default workspace skill loader saves initialization failure reasons across restarts. A bare
SkillLoaderkeeps them in memory. Persisting runtime plugin failures in quarantine requires a workspace store andrecord_declaration_failures=True.
After upgrading, run the audit from the project root and check which prerequisites it could measure.
bernstein govern audit --format json --strictGovernance and audit
bernstein govern audit writes text, JSON or SARIF 2.1.0. Run it from the repository root for workspace-relative SARIF locations. --quiet suppresses the report.
| Result | Default | --strict |
|---|---|---|
Any check fails or crashes, or a measured finding has passed=null |
1 | 1 |
| Otherwise, at least one prerequisite is unmeasurable | 1 | 2 |
| A nonempty selection returns only passes, declarations or both | 0 | 0 |
| Empty selection | 1 | 1 |
| CLI usage error | 2 | 2 |
The full table is in governance operations.
- Compare reports. Use
--since BASE_SHA256 --against HEAD_SHA256together to compare two stored audit reports offline. - Policy-library checks.
govern audit --only CMPruns the checks reported incompliance checkJSON. - ISO/IEC 42005 export.
bernstein audit export --standard iso-42005exports the evidence Bernstein holds for an AI impact assessment. - Control coverage.
bernstein compliance controlslists the benchmark suites that declare each control and their recorded measurements. Later submissions take precedence. The display saysunavailablewhen no measurement exists. - Scorecard baselines. An install-identity baseline stays neutral without a matching trusted key. Supply the public key with
bernstein bench run SUITE --baseline BASELINE.json --trusted-key FINGERPRINT=PATH. Stub baselines require--stub-signerand successful replay.
Evidence integrity and supply chain
- Delegation seals. Verification detects records removed, replaced or appended at the tail after close. It cannot detect replay of an authentic snapshot, an unrecorded action or changes made with the audit key. Older runs receive no backfill.
- Hash profile.
hash_profile: jcs-v2inbernstein.yamlopts a new run's journal, lineage spine and receipt into RFC 8785 canonical JSON. Bernstein signs the receipt only with a signing key configured.py-json-v1remains the default. - Release provenance. GitHub Releases include a Sigstore provenance bundle.
- OSV-Scanner. The adapter scans against a pinned local database and rejects digest mismatches and time-dependent rules (
ignoreUntil,effectiveUntil). This covers part of #2958. - CoSAI evidence map. The map includes MCP-T1 through MCP-T12 from the 2026-01-08 revision.
- Custody signers. Detached-JWS lineage signers can use an existing custody signer without exporting the private key.
- Offline signing. Offline wheelhouse signing keeps working with cosign 3.1.4.
Fleet routing and resources
- Dispatch tiers. Generated tasks get a dispatch tier from their runtime requirements. Each batch runs tasks from one tier. Tiers survive REST requests, PostgreSQL storage, YAML plan loading and both retry paths.
- Routing hints. The
model,effort,cliandrequireshints survive a restart in the REST API and the JSONL TaskStore. - Worker heartbeats. Disk, memory and mesh-latency readings can guide node selection, task stealing and worker pulls.
- Live view.
bernstein liveshows registered remote workers and the tasks they claimed.
Execution correctness
- Cost corrections. The budget stop counts increases in explicit cumulative cost above spend already recorded, even when token counts stay flat. The current snapshot format omits the prior explicit cost. After restart, the first report for a restored entry can miss a cost increase if no tokens were added.
- Orphaned tasks. After a clean worker exit, an eligible non-artifact task with no new commits or worktree changes can complete if its declared completion signals already pass on the base commit.
- Completion timeout.
BERNSTEIN_SIGNAL_TIMEOUT_Ssets the timeout for atest_passescommand. It defaults to 120 seconds. - Sandbox leases. A holder can release only its own active lease.
bernstein doctorreports owner task statuses for confirmed leaks older than the configured threshold (24 hours by default). It warns when a confirmed leak's age is unknown. - Reconcile helper. It skips completed work, stops on failure and can resume a trusted partial run.
- Required gates. A failing or inconclusive plugin result blocks promotion when
GatePipelineStep.requiredis true. An explicit block from the plugin also applies on an optional step.
Models and adapters
- OpenAI and Codex.
provider=openaiuses the OpenAI API client. The Codex adapter validates and forwards reasoning effort before spawning. - Retries. An explicit role or run model pin takes precedence. Otherwise, a retry keeps a supplied provider-prefixed model. With neither a known model nor a pin, the adapter resolves its default (#6605).
- Pi. Pi runs with
--print --mode jsonand streams events to the runner log. - New adapter contracts. The release adds CLI contracts for
cline,autohand,openhands,cody,composioandralphex. Adapters generated from the template match the currentCLIAdaptercontract.
MCP and spawn policy
- MCP tasks. The persistent TaskStore backs task creation, polling, result retrieval, listing and cancellation on the 2025-11-25 wire contract.
- Capability observations. With
--observe-capabilities, the gateway hashes each complete tool definition and each top-level field after the upstream handshake. The snapshot stores those digests. On a change,mcp.capability_driftrecords previous and current digests and the changed field names per tool. The repository includes test vectors. - Server manifest verification. With signing checks enabled and an audit chain supplied, the MCP pre-load check records the server manifest's signature verdict, the bundle scan status and the load decision.
- External policy registry. For fresh direct-adapter spawns, an injected registry checks each attempt's adapter, model and endpoint, including failover. At least one policy hook must return
ALLOW. - Role allow-lists. Roles without a named policy can use a
*adapter allow-list. An empty default denies every adapter.
What is partial in 4.0.0
| Component | Works in 4.0.0 | Not included |
|---|---|---|
| Graph trace ingest (#4964) | Maps recorded structure to RunPlan / PlanNode and hashes tool-call arguments |
Scheduling, runtime callback conformance, CLI, replay and offline receipts |
| Plan graph fold (#3614) | Rebuilds the recorded plan graph offline from the journal | Scheduler integration, amendments |
| Agent Manifest export | Composition data, signed availability companion and per-hop artifact references | A complete verdict |
| Registry federation | Opt-in for Python callers, Ed25519 pins, unsigned indexes refused by default | Donor CLI configuration |
| Attachment manifests | A read-only preview returns copies of selected existing tasks and the manifest digest | Creating or persisting derived tasks |
| External policy registry (#4912) | Checks fresh direct-adapter attempts, including failover | CLI policy loading and other execution paths |
| Worker telemetry | Node selection, task stealing, worker pulls | Orchestrator claims by task ID (#7853) |
| Registry partitions | Partitions inside one process | HTTP sharding |
| Operator halt controller | Pauses and resumes new TaskStore claims | Stopping running workers, installation-wide halt |
| Retry model retention (#6605) | Keeps a supplied provider-prefixed model unless a role or run pin overrides it | Resolving the adapter default in the retry request |
Dashboard, diagnostics and developer experience
- Dashboard. Task status shows tenant-scoped coordinator readiness. MCP status reports the coordinator and bridge versions separately.
- Quarantine. Workspace quarantine can also record reasons for malformed routine trigger declarations and invalid issue trackers.
- Windows. Artifact-mode completion handles agent IDs containing
:. A directory at a declared output path reports "wrote no output". - Telemetry.
bernstein telemetry verify-spanreads span JSON from standard input with--span -or--span @-. - Snapshots work inside linked Git worktrees. Flaky-test detection retains pytest errors, and GitLab webhook events follow configured trigger rules.
Documentation, CI and maintenance
- Workflow docs. The manifest reference distinguishes dispatch, verified completion and successful command execution. Resume examples supply the goal again because snapshots omit it.
- CI. Dependabot skips auto-merge on draft PRs. Maintainer review on protected automation changes excludes known writers. A maintainer PR follows ordinary review rules if a commit author is neither on the roster nor the built-in
github-actions[bot]account. - Characterization tests. Strict expected failures document known gaps in workflow execution and recovery (#6639).
Known limitations
- With an injected policy registry, bridge, in-process, container, sandbox and crash-resume paths refuse before execution. CLI policy loading is still pending (#4912).
- The detached local path without an SSH sidecar records task lifecycle events but does not yet execute the submitted task.
- Regular fixes ship on 4.0.x. The only planned 3.24.x backport is 3.24.2.
The complete list of changes follows below and is in docs/release-notes/v4.0.0.md.
Full list of changes
Added
- Validate attachment manifests and preview their derived tasks without writing to the task store.
(#6912) - Define an optional sandbox backend contract for verified reattachment and removal.
(#6911) - Include tenant-scoped coordinator readiness observations in dashboard task status.
(#7022) - Compare two stored governance audit reports offline with
govern audit --sinceand--against.
(#7029) - Let MCP gateways explicitly opt in to recording capability observations.
(#7036) - Export ISO/IEC 42005 impact-assessment evidence with
bernstein audit export --standard iso-42005.
(#7085) - Find the first changed or missing property in a pair of receipt render captures.
(#7101) - Aggregate supplied run-conduct projections within a specified time window.
(#7102) - Attach the Sigstore provenance bundle to GitHub Releases.
(#6968) - Provide an opt-in failure sentinel for testing audit detection, recording, and notification.
(#6916)
Changed
- Select fleet instances based on explicit worker capabilities rather than generic matches.
(#6910) - Update shipped Web UI dashboard assets with React Query 5.105.1.
(#6933) - Expose doctor installation checks as governance-audit report producers.
(#6918) - Migrate the quality-gate template to the standardized
GatePlugininterface.
(#6964) - Quarantine skills that fail initialization with a journaled reason instead of failing silently or
crashing. (#6917) - Capture MCP server capability snapshots only after completing the upstream handshake.
(#7016) - Report the coordinator version separately from the bridge version in MCP status.
(#7023) - Make the declaration quarantine check callable through the audit check registry.
(#7033) - Classify and assign dispatch tiers to generated tasks based on runtime requirements.
(#7034) - Derive reconciliation intent keys and check whether a fresh observation already satisfies the intent.
(#7040) - Cache snapshot metadata directory resolution once per listing for improved performance.
(#7048) - Report both implementation modules when a duplicate audit check is registered.
(#7051) - Update
agentsdependency to^0.33.0in the Cloudflare MCP server template.
(#6950)
Fixed
- Show registered remote worker nodes and their claimed tasks in the
bernstein liveAGENTS panel.
(#6753) - Route
provider=openairequests directly through the OpenAI API client.
(#6884) - Validate and forward reasoning effort parameters before spawning tasks in the Codex adapter.
(#6899) - Keep tasks from different dispatch tiers in separate execution batches.
(#6905) - Use the configured task server in role prompts, prompt partials, and the manager task-API skill.
(#6924) - Preserve task dispatch tiers across REST API requests and PostgreSQL persistence.
(#6908) - Make the completion-signal timeout configurable in place of the fixed 120-second limit.
(#6923) - Build a verified removal proposal for one leaked sandbox lifetime.
(#6909) - Record MCP tool-signing gate outcomes in the audit chain.
(#6913) - Include full injected tool definitions per server when calculating spawn prompt token budgets.
(#6907) - Fix datasource query receipt recording on Windows by sanitizing directory paths containing colons in
agent IDs. (#6928) - Omit unsupported sampling parameters when the OpenAI Agents adapter uses GPT-5 mini with the Responses
API. (#6975) - Complete clean orphaned tasks when their verification gates already pass on the base commit.
(#6900) - Ensure standard input streams are read reliably across platforms using Click
open_filein telemetry
commands. (#6989) - Prevent releasing other active leases by ensuring sandbox holders release only their own assigned
leases. (#7001) - Let the pipeline enforce the
requiredflag for failures reported by a quality-gate plugin.
(#7003) - Quarantine malformed routine trigger declarations to isolate configuration errors without halting
execution. (#7007) - Journal quarantine events for invalid issue trackers discovered during workspace scanning.
(#7008) - Automatically quarantine runtime plugins that encounter initialization failures.
(#7009) - Apply configured trigger rules when processing incoming GitLab webhook events mapped to tasks.
(#7010) - Preserve MCP capability snapshots when concurrent writers update the same store.
(#7026) - Preserve explicitly declared dispatch tiers defined in YAML plan steps.
(#7027) - Verify that the inventory anchor names the inventory recorded by a governance audit report.
(#7038) - Fix snapshot resolution and creation when operating inside linked Git worktrees.
(#7045) - Reject duplicate tunnel driver registrations to prevent driver conflicts.
(#7046) - Handle shallow registration call stacks gracefully in registry guards without raising inspection
errors. (#7047) - Recover the run outcome before recording the run-quiescence observation.
(#7049) - Preserve pytest error outcomes when parsing test evidence in flaky test detection.
(#7025) - Update vulnerable glob and minimatch dependencies and hash-pin the affected pip installs.
(#6965) - Correct documented CLI commands, flags, and the benchmark Lambda link.
(#7079)
OSV-Scanner adapter, pinned to a local database
get_scanner("osv-scanner") returns a feed-pinned SCA adapter. It runs
osv-scanner scan source --offline against a local OSV database
(OSV_SCANNER_LOCAL_DB_CACHE_DIRECTORY), never against osv.dev, whose answer
changes as advisories are published. The caller pins the database in
scope.config["db_pin"]. The adapter digests every osv-scanner/<ecosystem>/all.zip
under the root and refuses to scan when the digest differs, as the Trivy
adapter does for trivy.db.
Finding identity includes the vulnerability id, normalized lockfile path,
severity, summary and affected package metadata. Windows drive and UNC file
URIs use the shared path normalizer so checkout locations do not change
in-scope finding hashes. Any osv-scanner.toml ignore list in the target tree
feeds the recorded invocation hash. Include/exclude globs and max-depth scope
limits are refused rather than ignored. Partial implementation of #2958.
SARIF reports that explicitly mark any invocation as unsuccessful are
rejected, including reports with no findings. Absent invocation metadata
and successful invocations preserve the existing finding normalization.
Config identity uses the versioned osv-scanner-config-v2 digest of sorted
target-relative POSIX paths and exact file bytes: moving an ignore list between
package directories changes invocation identity, while moving the entire checkout
does not. Before any vendor process, the adapter parses TOML and rejects explicit
IgnoredVulns.ignoreUntil and PackageOverrides.effectiveUntil fields (including
case variants), because OSV-Scanner evaluates them against the current time.
Timeless configs and temporal words in comments or reasons remain supported;
unreadable or invalid TOML configs are refused rather than left unvalidated.
Doctor environment snapshots remain statically typed
Doctor's adapter lookup narrows an optional environment mapping where it reads
PATH, clearing an advisory mypy error without an ignore or a cast. Explicit
paths, including an empty PATH, and fallback to the ambient environment retain
their existing behavior.
Fixed
- Qualify the catch-up skip list annotation so the schedule store's
list
method does not shadow the built-in type during mypy checks.
Derive the recorded structural plan graph offline
A pure journal fold, exported as bernstein.core.replay.fold_plan_graph, now
reconstructs structural task nodes using the producer's
canonical graph encoding. Missing latest snapshots return an explicit
unavailable result, and malformed graph evidence fails closed. Generated
producer-prefix tests cover independent append failures and determinism across
processes. Scheduler integration and amendments remain deferred (#3614).
Opt-in registry federation for Python callers
Volunteer registry callers can explicitly enable bounded breadth-first index
resolution with per-URL Ed25519 pins, default unsigned refusal, denial rules,
cycle diagnostics and conflict reporting. Project manifests remain the policy
source. Existing flat-index browse behavior is preserved; donor CLI
configuration for federation remains a follow-up.
Fixed
- Programmatic spawn admission now consults an injected external policy registry before starting an action, records every hook evaluation in the audit chain, and refuses unavailable engines, missing explicit approval and unsupported obligations. Native admission refusals remain authoritative. CLI policy loading remains a separate integration task. (#4912)
- Direct-adapter attempts, including failover, are admitted against their actual adapter, model and effective endpoint. Policy or audit refusals cannot trigger another executor. Injected registries refuse unbound bridge, in-process, container/sandbox and crash-resume paths; native-only behavior is unchanged.
Graph trace ingest adapter for graph-shaped agent runtimes
Ingests foreign execution traces from graph-shaped agent runtimes through the ingest adapter plugin boundary without modifying core scheduling logic (#4964).
The adapter retains foreign nodes, edges, input/output state and tool calls in its ingested representation, and maps node identities and edges to Bernstein's lineage task graph vocabulary (RunPlan / PlanNode). Every tool call carries a deterministic digest of its normalized arguments. Graph text renderings remain byte-identical across repeated ingests.
Input/output state must be mappings, edge conditions must be booleans, and node
identifiers and edge endpoints must be non-empty strings. Tool calls require a
non-empty name. Display fields reject control characters and line separators
before rendering. Missing roles remain unassigned. The lineage projection keeps
cycles and flattens conditional edges into dependencies; it describes recorded
structure and is not a schedulable execution plan.
This is the entry-point-registered ingest-boundary slice of #4964. String inputs
are JSON; local files require an explicit Path. Malformed tool-call entries are
rejected instead of silently discarded. The fixture is synthetic, and the
dependency-free callback recorder is tested with hand-driven events. Real-runtime
callback conformance and recorded-fixture provenance, production declaration
collection, CLI/replay wiring, and offline receipts remain follow-up work.
- Allow reusable lineage detached-JWS signers to use an existing custody signer without exporting private key material. Existing PEM callers and signature bytes remain compatible.
Expose all existing CMP policy-library observations through govern audit --only CMP and include identical workspace findings in compliance check
JSON. Policy snapshots now use project observations unless explicitly
overridden by declaration flags. Framework profiles mark required check IDs
without changing verdicts or asserting conformance.
CMP execution errors retain their check area and use the shared registry's
exception isolation on both JSON surfaces. Under govern audit --strict, a
crashed CMP check returns exit 1 rather than the incompleteness exit 2.
The existing framework profile annotation is also retained in each SARIF
result's properties, without changing the check verdict or command exit code.
Governance audit JSON, SARIF and CI exit codes
bernstein govern audit now supports --format text|json|sarif, --strict
and --quiet. JSON preserves the complete finding fields, including id and
message (also included in doctor --json). SARIF 2.1.0 includes check rules
and workspace-relative file locations; passing checks use kind: pass and
level: none. Unmeasurable results use kind: review and level: none, with
their reason in the message. Evidence locators are omitted, and absolute local
paths in diagnostics are redacted. Run from the repository root for code scanning.
--quiet suppresses reports and check-crash warnings while retaining sentinel
journaling and notifications.
The default gate keeps exit 1 for failures, check crashes and unmeasurable
checks. --strict distinguishes unmeasurable prerequisites with exit 2,
unless a check failed or crashed (exit 1 takes precedence). Measured findings
without an explicit pass fail closed. Click usage errors, including incompatible
output flags, also exit 2. Declared findings no longer fail the gate: they
previously caused exit 1 on main and now allow exit 0 when no other finding
blocks it. An empty selection exits 1. See the complete exit-code table in
governance operations.
Add a callback-based reconcile apply helper that records already-satisfied intents as skips, stops on failure, and resumes trusted same-context partial receipts without repeating successful changes. Retain stable intent keys and require separate policy authorization.
-
Add one
hash_profileselection inbernstein.yamlfor a new run's journal, lineage spine and receipt (signed when a signing key is configured).jcs-v2opts into RFC 8785; unknown profiles fail closed. -
The first release containing this fragment is the declared compatibility release.
py-json-v1remains the default for that whole release. Changing the new-run default tojcs-v2requires a separate later PR after measured soak; this release does not flip it. -
Existing evidence retains its recorded profile, or
py-json-v1when absent, and reopens and verifies without rewriting historical bytes. -
Preserve declared verification scopes in plugin results and cached gate reports, and retain the reason for inconclusive plugin results. Legacy cache entries remain without coverage declarations.
Delegation verification detects changes to a sealed final receipt chain
Run close atomically seals the final delegation digest, hop count and manifest
configuration with the install audit key. Every delegation verifier checks that
seal and the declared head/count: tail deletion, substitution and post-close
growth fail, even if a keyless writer recomputes the manifest checksum. Missing
declarations are incomplete (valid=false, exit 3), including pre-upgrade runs
and interrupted finalization. Existing unsigned head declarations fail seal
verification; they are not silently upgraded from the remaining ledger (#5352).
Open and legacy runs retain their intact, authority-checked delegated ceilings
in the agent registry, marked by incomplete_runs; saved version-1 registry
snapshots remain verifiable. Access reviews exclude incomplete or mismatched
chains from verified sources. Incomplete chains never report a passing verdict.
This does not detect replay of an earlier authentic snapshot, prove that every
runtime action was recorded, or protect against a writer holding the audit key.
Deleting a declaration yields incomplete, never a completeness claim. There is
no automatic backfill: the remaining receipts cannot establish a historical tail.
Compliance control-to-suite coverage and latest measurements
bernstein compliance controls now surfaces the declaring benchmark suites and
latest recorded measurement for each registry control across text, Markdown,
and JSON outputs (#5455, #7098):
- Built-in benchmark suite control declarations are exposed from a single shared
mapping inbernstein.eval.bench.bench_cli, paired with tests verifying every
registered built-in suite declares at least one valid registry control ID. bernstein.compliance.control_coveragejoins the central control registry
with built-in (and optional custom.suite.json) declarations and
SubmissionBundlerecords, selecting the latest measurement per suite
deterministically regardless of input order and reportingunavailablewhen no
measurement exists.bernstein compliance controlsaccepts optional repeatable--bundleand
--bundle-dirflags (falling back to.sdd/bench/.sdd/benchmarkswhen
omitted) while preserving existing control keys and stable control ordering.
Authenticate install-identity benchmark baselines in scorecards
Python callers can pass trusted public keys to evaluate_ci_scorecard so a valid
install-identity baseline can produce a verified success or regression result.
Missing trust, invalid signatures and failed receipt replay remain neutral.
bench run --baseline accepts --trusted-key FINGERPRINT=PATH to supply
those public keys explicitly. Without a matching key the comparison stays
neutral.
Keep the pytest controller outside the memory cap
The test suite applies its address-space guard to test workers and plain
runs, leaving the pytest-xdist controller uncapped. Resource failures and
the reported worker-crash controller abort include guard context. Diagnostic
regressions run without the POSIX-only probe on Windows and preserve custom
caps and inherited Python module paths in child sessions.
Add optional composition-only Agent Manifest export with operator declarations,
a signed hash-bound availability companion, and per-hop artifact references.
Pin agent-manifest 0.12.0; preserve its INCOMPLETE composition-only verdict and
strict runtime artifact verification semantics.
CoSAI packs include the pinned MCP threat taxonomy (#6217)
The existing cosai evidence map adds MCP-T1 through MCP-T12 from the selected
2026-01-08 paper revision. Every new row preserves its exact source provenance
and names bounded gate evidence and remaining gaps: four Partial and eight
Not covered, with no claim that the pre-load gate covers an entire threat class.
The original thirteen principle rows, pack schema and ZIP layout are preserved.
Worker health telemetry
Worker heartbeats accept optional disk, memory and mesh latency readings.
The registry uses them for direct node selection, task stealing and
bernstein worker pulls. Orchestrator claims by task ID remain open in
#7853.
Cluster registry ring
Connect the deterministic hash ring to opt-in local partitions inside one
registry process through cluster.shards and cluster.shard_vnodes. Node
registration, heartbeat and administrative mutations use the owning partition;
defaults and the flat node persistence format remain unchanged. HTTP sharding is not enabled.
Fixed
- Preserve backend dispatch tiers through both task retry paths. Reopen can explicitly select a validated tier, recording the previous tier, selected tier and reason with the persisted transition; omission preserves the current tier. Automatic failure-based retiering remains outside this change. Refs #6560.
Preserve task routing hints across restart
Task REST responses and the JSONL TaskStore now retain exact model, effort,
cli and requires declarations through single/batch creation, later records
and restart. Invalid capability declarations are rejected before REST
persistence; historical tasks keep their existing defaults. This preserves
routing data without adding fleet selection, dispatch or result acceptance.
Refs #6562.
Doctor reports aged leaked lanes with their evidence (#6572)
The existing read-only leak check now includes each confirmed lane's recorded
owner statuses and resource-age source in text and structured JSON. Both sweep
kinds use one canonical task replay snapshot. The existing 24-hour environment
default and strict older-than comparison remain; unusable ages warn explicitly.
Unreadable evidence or an invalid threshold fails the doctor check instead of
being displayed as a successful check. This adds no removal authority or
provider probe.
Retain supplied provider models across task retries
Both retry task creators preserve an already supplied provider-prefixed model
instead of manufacturing a bare Claude tier. Explicit role/run model pins keep
their priority, and existing effort, retry limits and backoff remain unchanged.
This is a partial fix for #6605: missing effective model context still needs
propagation when neither the task nor an explicit pin identifies the model.
Keep the model override unset in both retry paths when neither a prior task
model nor an operator pin is available. The adapter continues resolving its
configured default, rather than receiving an invented Claude tier. Original
effort overrides are preserved alongside an unresolved model so an effort-only
override cannot bypass the original template resolution. Known-model effort
escalation, explicit pins and resolved provider-model retention remain unchanged.
The seed parser now reads every quality_gates.* key that QualityGatesConfig defines. mutation_testing, dep_audit, intent_verification, large_file_check, migration_reversibility_check, test_expansion, agent_test_mutation, integration_test_gen, review_rubric and their sub-keys were documented and accepted by the schema but never forwarded, so setting them in bernstein.yaml left the gate at its default (#6615). Two consequences for existing seeds: this repository's own bernstein.yaml sets large_file_check: true, which now takes effect, and a quality_gates.intent_verification: that is not a mapping, or an integer key such as timeout_s given a bare true, now raises SeedError instead of being ignored.
Preserve durable MCP tasks on the 2.3 SDK
Upgrade the MCP dependency to the tested 2.3 release line. Keep the existing
2025-11-25 task-create, polling, result, list and cancellation wire contract
backed by Bernstein's authenticated persistent TaskStore, despite removal of
the SDK's experimental task runtime. Ordinary SDK2 clients retain normal tool
calls. This does not implement or advertise the redesigned Tasks extension.
The compatibility adapter uses the SDK's provisional middleware API, so the
dependency is bounded below 2.4 until another compatibility review. Task
cancellation now supplies the existing HTTP endpoint's required JSON body.
The custom HTTP transport retains its four implemented handshake-era protocol
revisions. SDK2 clients in automatic mode negotiate the existing 2025-11-25
wire format; explicitly requested modern HTTP revisions are rejected until
that transport implements their response contract.
Update the Kubernetes client to 37.0.1
Update the Kubernetes extra and its required dependencies. Scope the dependency
review exception for typing-extensions 4.16.0 to its published PSF-2.0 license
metadata, which is already accepted by the project (#6621).
Workflow completion and authority characterization
Adds focused local-process characterization for workflow completion and authority.
Strict expected failures record the proposed completion gaps without changing
runtime behavior; positive controls and process cleanup remain blocking (#6639).
Workflow dispatch crash and restart characterization
Adds focused local-process characterization for workflow dispatch crash and restart.
Strict expected failures record the proposed completion gaps without changing
runtime behavior; positive controls and process cleanup remain blocking (#6639).
dependabot-auto-merge.yml no longer tries to arm auto-merge on a Draft update, which the API refused and which failed the lane; a Draft finishes without calling gh, the new ready_for_review trigger brings it back once it is marked ready, and a Draft state the event did not report as true/false fails closed (#6758).
Security
- Required maintainer reviews on protected automation and workflow PRs now
exclude known writers, preserving the existing proven unchanged-diff
exception and normal governance gates. The same exclusion applies to the
additional maintainer review for large ordinary changes.
Ship the React Query 5.105.3 client in the dashboard
Update React Query and its core package to 5.105.3, rebuild the bundled
dashboard, and refresh its documented screen captures and bundle binding
together (#6941).
Pi agents stream progress to the runner log
Pi now runs with --print --mode json and receives EOF on stdin, so it
streams task events to the runner log and exits after the prompt instead
of opening its interactive UI or waiting on an inherited stdin pipe.
Log growth supplies the existing liveness signal, preventing heartbeat
escalation from treating an actively streaming agent as idle before its
first artifact (#6943).
Generated adapters match the current CLIAdapter contract
Adapters created from the cookiecutter template now construct successfully
and accept the current spawn arguments. They launch asynchronously and return
the live process handle, PID and log path, with timeout watchdog support.
Generated tests cover the current contract instead of removed result fields
(#6976).
Security
-
Extend the gateway's opt-in initialization capability observation to fingerprint
complete parsed tool definitions and record per-tool changed fields. Preserve
legacy name-only audit digests and migrate snapshots without inventing a prior
definition baseline. This observation does not enforce call holds or approvals
(refs #6984). -
Add an explicitly configured persistent operator controller for new TaskStore claims, with fail-closed state validation and serialized halt/resume. This first slice does not stop running workers or implement installation-wide halt.
Type remote status and cancellation observations explicitly
RuntimeBridge now offers optional typed remote status and cancellation
capabilities that refuse unsupported calls without invoking legacy methods.
Frozen DTOs distinguish unknown and waiting states from observed terminal
lifecycle, and unconfirmed cancellation from confirmed cancellation. This
partial component preserves existing bridges; runtime integration and the
observed-event journal remain separate work (#6992).
Added
- Support an optional
*adapter allow-list for roles without a named policy.
An empty default denies every adapter; existing named-role behavior and
policies without a default retain their semantics (#7019).
Fixed
- Apply ordinary contribution review requirements to maintainer-opened pull
requests carrying commits authored outside the current roster. Co-author
credit alone preserves the existing rule (#7020).
Preserve offline wheelhouse signing with cosign 3.1.4
Probe offline compatibility flags by parsing them with --help, so flags
hidden from the displayed help still apply. The signing script and real
airgap tests retain the existing detached signatures and offline behavior
when upgrading cosign to 3.1.4 (#7044).
Ship the React 19.3 runtime with the matching browser bundle
Update React and React DOM to 19.3.1 and their type packages to 19.3.0.
Rebuild the bundled dashboard and recapture its documented screens so the
wheel ships the dependency versions recorded in the lockfile (#7052).
- Use the registered adapter key for admission checks when dispatch falls back to the run pin. Pinned Pi now looks up the existing
picontract. Admission still requires passing evidence and a matching receipt.
Require both trufflehog dependency updates before Renovate opens their grouped branch, and check image extraction and group configuration alongside the existing scanner version gate.
Add an optional bounded render-probe capability to the pinned Playwright driver, with engine-computed styles, checked environment signals, and explicit unavailable results for missing evidence or incomplete captures.
Render comparisons retain unstable properties from both receipts
Render receipt comparisons retain each side's unstable properties, including
unchanged values and incomparable pairs. Existing environment checks and
comparison verdicts are unchanged (#7088).
Show unreadable runs in conduct period summaries (#7093)
Add UnreadableRun and UnreadableRunReason (absent, unverifiable,
torn) to ConductAggregate and aggregate_conduct so supplied run IDs whose
journals could not be read appear in deterministic order alongside valid runs
with unverified coverage, keeping the period coverage unverified.
Conduct aggregates can compare caller-supplied recorded activity scopes against
an effective period scope using the existing authority rules. The read-only
section names divergent axes with source run/activity/event ranges and shows
missing evidence as unavailable or partial.
Goal-drift benchmarks can report recorded intent calibration policy/version,
sample count and existing quality metrics. Receipt replay rederives the report
without fitting a model or changing hard-check results; missing or ambiguous
calibration is explicitly unavailable.
Document the ArtifactSink storage contract (#7169)
Add a short protocol reference for ArtifactSink to
docs/operations/artifacts.md, covering logical key naming, current async
method signatures (write, read, list, delete, exists, stat,
close), lexicographically sorted list() results, and the durable=True
write guarantee.
Cleanup force and inspection are documented separately
Explain selected worktree and branch removal, the additional branch pruning
permitted by bernstein cleanup --force, and best-effort preservation. Show
read-only worktree inspection before cleanup (#7301).
Clarify product and workflow documentation (#7308, #7311, #7312, #7314, #7316, #7318)
The visible homepage, evaluation guide, adapter scope and present-day blog
footers describe Bernstein as an open-source governance layer for AI agents.
The historical ACP entry names the integration it shipped. The workflow
manifest reference distinguishes successful agent dispatch from verified
completion and from successful command execution.
Preserve goal substitution in workflow resume examples
The workflow guide passes the original goal in both resume examples, including
the explicit-manifest form, and lists the saved manifest snapshot fields.
It explains that resume defaults to an empty goal and does not restore it from
the snapshot (#7317).
Clarify README dashboard connections to an existing task server, identify administrative MCP tools and returned run identifiers, and make the name analogy specific to orchestration. Keep all translated README sections aligned.
Artifact-mode completion works on Windows. Its signing identity's Agent Card was written to .sdd/agents/agent:artifact-completion/, and : is not a legal Windows path component, so every artifact-mode task failed there with NotADirectoryError. The card now lives under the SHA-256 of the agent id, the encoding datasource receipt cards already use; the lineage gate matches cards by content, so verification is unchanged. A directory at a task's declared output path now reports "wrote no output" on Windows too, instead of a raw PermissionError.
Contracts for the cline, autohand, openhands, cody, composio and ralphex adapters
Admission refuses an adapter that has no contract. cline, autohand,
openhands, cody, composio and ralphex now have one in tests/contract/contracts/, pinning the flags
each adapter always passes: cline --yolo; autohand --unrestricted --prompt; openhands --headless --override-with-envs --task; cody chat --message;
ao spawn --prompt; ralphex --no-color --task-model. Each was
checked with bernstein adapters contract-check against the current release
of its CLI (cline 3.0.72, autohand 0.9.8, OpenHands CLI 1.16.0, cody 5.5.26,
@aoagents/ao, ralphex v1.7.0).
Count explicit cumulative cost increases accurately
CostTracker.record_cumulative now charges the increase in an explicit
cumulative cost, including corrections with unchanged token counts. Duplicate
or lower reports do not add spend or refund it. Reports without an explicit
cost price each new token once and count toward the next explicit total for
that agent, task and model. New tokens appear in usage reports even when
their monetary delta is zero. Recorded spend drives the existing budget stop.
The persisted format is unchanged. For a key restored from a snapshot, its
first explicit report keeps the previous token-based proration or duplicate
suppression and seeds the monetary baseline for later reports. Estimates
already recorded since the restart count toward that first explicit amount. Exact cost
corrections spanning a restart cannot be recovered from those snapshots.
Fail detached SSH tasks on a nonzero remote exit
Detached SSH execution now records task.failed with the remote exit code
when a task command fails, instead of recording task.completed. The SSH
receipt retains the exit code, and later tasks stay pending. Both the supervisor
and the SSH goal helper close the run as failed. Restarting the supervisor
restores a failed result if it was interrupted after the journal closed.
Concurrent finalizers reuse the first journal outcome and its terminal receipt.
The known limitations also clarify that the
detached local path without an SSH sidecar records task lifecycle events but
does not yet execute the submitted task.
Update reviewdog/action-actionlint to v1.83.0
The Workflow lint job now uses reviewdog/action-actionlint v1.83.0. It keeps the existing reviewdog reporter, error threshold and ShellCheck setting.
Update peter-evans/create-pull-request to v8.1.2
The topology healer, weekly and continuous coverage ratchets, observability snapshot and major/minor release bump use create-pull-request v8.1.2. Their branch names, token selection, inputs and permissions stay as configured. This action uses Node.js 24 on the existing Ubuntu runners.
Update crate-ci/typos to v1.52.0
The spelling job and deterministic auto-heal installation use the same typos v1.52.0 commit. The spelling scan keeps its default read-only mode, and auto-heal retains its strategy condition and explicit write_changes: false input.
A failed workflow node now stops every later layer
bernstein workflow run (and WorkflowRunner.run) used to keep
dispatching nodes in later layers after a node failed, whenever those
nodes depended only on branches that had passed. Those nodes ran their
commands or agents, wrote success checkpoints, and appeared twice in the
result. A failure now skips every later layer once, without dispatching
anything, matching bernstein workflow resume.
Known limitations
- #4912 (partial): CLI policy loading remains a separate integration task. Injected registries refuse unbound bridge, in-process, container/sandbox and crash-resume paths.
- Regular fixes ship on 4.0.x. The only planned 3.24.x backport is 3.24.2.
What's Changed
- fix(deps): update dependency agents to ^0.32.0 by @renovate[bot] in #6881
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6877
- fix(llm): route provider=openai through the API client by @chernistry in #6884
- docs: refresh docs/requirements.txt transitive pins by @aarav570 in #6803
- chore(ci): ratchet coverage baseline up to 85.2% by @chernistry in #6879
- feat(volunteer): resolve federated indexes with explicit trust and limits by @chernistry in #6860
- build(deps): raise pytest-benchmark floor to >=5.3.1 for pytest 9.2 (unblocks #6862) by @chernistry in #6898
- fix(codex-adapter): forward and validate reasoning effort before spawn (#6641) by @chernistry in #6899
- fix(orchestrator): never merge tasks with different dispatch tiers (#6888) by @chernistry in #6905
- fix: toolcall interlock: write the source-rights evidence design note (slice of #6409) (#6889) by @chernistry in #6906
- fix: fleet: select instances by explicit worker capabilities (slice of #6562) (#6892) by @chernistry in #6910
- fix: compliance: document ISO 42005 evidence boundaries and the role of ISO 42006 (slice of #6222) (#6896) by @chernistry in #6914
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6904
- docs(trace): correct the conformance command name and runtime measurement wording by @chernistry in #6885
- chore(deps): bump oauthlib from 4.0.0 to 4.0.1 by @chernistry in #6897
- fix: templates: 8052 is still hardcoded in role task prompts, two prompt partials and the manager task-API skill (#6299) by @chernistry in #6924
- fix: tasks: preserve dispatch_tier through REST and persistence (slice of #6560) (#6891) by @chernistry in #6908
- fix: Completion signals are killed at a hardcoded 120s, turning any slower check into an infinite retry loop (#6119) by @chernistry in #6923
- fix: govern audit: an injectable failure sentinel proves the detect, record, notify path end to end (#5091) by @chernistry in #6916
- fix: sandbox leaks: build a verified removal proposal for one leaked lifetime (slice of #6571) (#6886) by @chernistry in #6909
- fix: mcp: record signing gate outcomes in the audit chain (slice of #6217) (#6895) by @chernistry in #6913
- fix: CI: Speed up unit test runtime by increasing shards and parallel workers (#4756) by @chernistry in #6926
- docs(ci): name the repository ruleset that requires the quorum context by @chernistry in #6852
- fix: spawn budget: count the full size of injected tool definitions per server (slice of #6233) (#6887) by @chernistry in #6907
- chore(deps): update dependency platformdirs to v4.14.2 by @renovate[bot] in #6925
- chore(deps): update gitea/gitea docker tag to v28.3 by @renovate[bot] in #6932
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6930
- fix: datasources: query receipts cannot be recorded on Windows — agent_id with a colon is used as a directory name (#5859) by @chernistry in #6928
- build: refresh shipped SPA for React Query 5.105.1 by @chernistry in #6933
- chore(deps): update dependency platformdirs to v4.14.3 by @renovate[bot] in #6934
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6938
- chore(deps): update dependency @vitejs/plugin-react to v6.2.1 by @renovate[bot] in #6942
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6949
- fix(config): read the remaining seventeen quality_gates keys from the seed (#6615) by @Chirag6722 in #6831
- fix: preserve OSV report identity and failed invocation state by @chernistry in #6732
- fix(doctor): narrow optional environment before reading PATH (#2980) by @chernistry in #6703
- fix(deps): update dependency agents to ^0.33.0 by @renovate[bot] in #6950
- fix(ci): dependabot auto-merge leaves a Draft unarmed and fails closed on an unknown state (#6758) by @Chirag6722 in #6832
- fix: govern audit: every doctor check becomes a producer (#5073) by @chernistry in #6918
- chore(deps): update dependency astral-sh/uv to v0.13.7 by @renovate[bot] in #6946
- feat(security): enforce external policy on resolved programmatic dispatch by @chernistry in #6956
- chore(deps): update dependency backrefs to v8.2 by @renovate[bot] in #6955
- fix(workflows): stop dispatching later layers after a node fails by @chernistry in #6844
- docs(playbooks): drop the download figure from the README stats line by @chernistry in #6960
- docs: regenerate adapter last-green table from canary receipts by @chernistry in #6959
- chore(deps): update dependency platformdirs to v4.14.4 by @renovate[bot] in #6957
- docs(mcp): propose HTTP catalog boundary (#6456) by @chernistry in #6750
- docs(security): propose authenticated request bucket boundary (#6609) by @chernistry in #6698
- docs(receipts): define remaining reported coverage contract by @chernistry in #6701
- docs(readme): give the header logo its own background by @chernistry in #6961
- docs(mcp): propose execution readiness contract (#6645) by @chernistry in #6695
- docs(worker): propose remote claim base and identity contract (#6538) by @chernistry in #6697
- docs(doctor): propose aged leak finding boundary (#6572) by @chernistry in #6751
- docs(workflows): propose verified agent completion boundary (#6639) by @chernistry in #6696
- chore(deps): patch glob/minimatch advisories and hash-pin pip installs by @chernistry in #6965
- ci(codeql): analyse every pull request into main by @chernistry in #6966
- docs(mentions): add TRACE annex, Enterprise DNA, SkillsLLM by @chernistry in #6958
- chore(deps): bump huggingface-hub from 1.31.0 to 2.3.0 by @dependabot[bot] in #6620
- fix: conformance vectors: policy and approval — which rule, which version, whose sign-off (#5060) by @chernistry in #6927
- fix: OrcaRouter provider support for Bernstein (as an OpenAI-compatible endpoint) (#5546) by @chernistry in #6920
- fix(tasks): retain resolved provider models across retry paths (#6605) by @chernistry in #6745
- fix(tests): scope memory limits to pytest workers and preserve guard context by @chernistry in #6711
- fix(bench): verify trusted install signatures in scorecards by @chernistry in #6702
- docs(contributing): state that new functionality comes with tests by @chernistry in #6969
- chore(deps): update dependency ovsx to v1.3.0 by @renovate[bot] in #6693
- fix(adapters): omit unsupported sampling for GPT-5 mini responses by @chernistry in #6975
- chore(deps): update dependency hatchling to v1.33.0 by @renovate[bot] in #6978
- chore(deps): update dependency @types/node to v24.13.6 by @renovate[bot] in #6262
- docs: start the README header with the banner, single-row badge strip by @chernistry in #6963
- fix(ci): skip every scanner fixture target in the Trivy filesystem scan by @chernistry in #6974
- fix(orchestrator): complete an orphaned task when its gates already pass on base (#6606) by @chernistry in #6900
- chore(deps): bump the python-minor-and-patch group across 1 directory with 42 updates by @dependabot[bot] in #6937
- ci(release): attach a Sigstore bundle to GitHub Releases by @chernistry in #6968
- chore(deps): update dependency @types/node to v24.19.2 by @renovate[bot] in #6980
- fix(typing): distinguish planned and recorded crew receipt bodies by @chernistry in #6981
- chore(deps): update dependency @types/vscode to v1.145.0 by @renovate[bot] in #6982
- test(ticket): generate only undeclared metadata keys by @chernistry in #6983
- fix(telemetry): read stdin through Click open_file by @chernistry in #6989
- fix(templates): build the quality-gate template on GatePlugin by @chernistry in #6964
- fix(typing): type the probe collector callback by @chernistry in #6996
- fix(typing): narrow benchmark baseline case identifiers by @chernistry in #6997
- fix: orchestration: add strict attachment manifest validation and read-only preview (slice of #6647) (#6894) by @chernistry in #6912
- fix: sandbox: add the optional removal backend contract (slice of #6571) (#6893) by @chernistry in #6911
- fix(skills): quarantine a skill that fails to initialise, with a journaled reason (#5108) by @chernistry in #6917
- chore(deps): update dependency platformdirs to v4.14.5 by @renovate[bot] in #7004
- fix(typing): declare core compatibility import loaders by @chernistry in #7000
- fix(sandbox): release only the lease this holder owns by @chernistry in #7001
- docs: remove the funding file and the links to the sponsors listing by @chernistry in #6999
- fix: govern audit: JSON and SARIF output, and an exit code CI can gate on (#5076) by @chernistry in #6915
- fix(templates): make the quality-gate template follow the pipeline's required flag by @chernistry in #7003
- feat(adapters): ingest adapter for graph-shaped runtime traces (#4964) by @chernistry in #7005
- fix(triggers): quarantine malformed routine declarations by @chernistry in #7007
- fix(trackers): journal tracker quarantine in workspace discovery by @chernistry in #7008
- fix(plugins): quarantine failed runtime plugin initialization by @chernistry in #7009
- fix(webhooks): apply trigger rules to mapped GitLab tasks by @chernistry in #7010
- Snapshot MCP capabilities after the upstream handshake by @chernistry in #7016
- Preserve MCP bridge liveness during coordinator connection errors by @chernistry in #7017
- Expose scoped coordinator observations in task status by @chernistry in #7022
- Project coordinator version separately in MCP status by @chernistry in #7023
- Preserve pytest error outcomes when parsing test evidence by @chernistry in #7025
- chore(deps): update dependency lucide-react to v1.67.0 by @renovate[bot] in #6948
- Compare stored governance audit reports offline by @chernistry in #7029
- Preserve concurrent MCP capability snapshots by @chernistry in #7026
- Preserve explicit dispatch tiers in YAML plan steps by @chernistry in #7027
- Implement the callable contract for the quarantine check by @chernistry in #7033
- Classify dispatch tiers for generated tasks by @chernistry in #7034
- Record structured pytest phase observations by @chernistry in #7035
- Enable explicit MCP gateway capability observations by @chernistry in #7036
- chore(deps): update dependency lucide-react to v1.68.0 by @renovate[bot] in #7032
- Verify the inventory identity named by an audit report by @chernistry in #7038
- Derive reconcile intent keys and check fresh observations by @chernistry in #7040
- Accept trusted baseline public keys in bench run by @chernistry in #7041
- Evaluate stable structured fix proof observations by @chernistry in #7042
- Fix snapshots in linked Git worktrees by @chernistry in #7045
- Reject duplicate tunnel driver registrations by @chernistry in #7046
- Handle shallow registration call stacks by @chernistry in #7047
- Resolve snapshot metadata directory once per listing by @chernistry in #7048
- Recover the outcome before the run quiescence observation by @chernistry in #7049
- Update sigstore/cosign-installer action to v4.1.3 by @renovate[bot] in #7014
- Update zizmorcore/zizmor-action action to v0.8.1 by @renovate[bot] in #6945
- fix(artifacts): artifact completion works on Windows by @Chirag6722 in #6952
- docs(sdd): the LineageGate projection does not yet check the journal head (#7013) by @chernistry in #7015
- docs(skills): credit #5750 for the declaration quarantine store design by @chernistry in #7011
- Name both implementations of a duplicate audit check by @chernistry in #7051
- fix: delegation: removing the final receipt leaves a shorter chain that still verifies (#5352) by @chernistry in #6921
- fix: Leaked lanes: a sweep lists worktrees and sandboxes whose task is terminal, and removes them behind the removal gate (#5112) by @chernistry in #6919
- docs(quality): refresh the context file against its subtree (#6939) by @vaibhav8a in #6967
- ci(trufflehog): move the action pin and image version to 3.99.8 together by @chernistry in #7058
- Add web UI regression harness by @chernistry in #7031
- chore(webui): neutral package names in the render seed list by @chernistry in #7006
- docs: correct CLI examples and benchmark anchor by @chernistry in #7079
- Bernstein issue #3527 by @chernistry in #7084
- Bernstein issue #6985 by @chernistry in #7083
- Bernstein issue #6222 by @chernistry in #7085
- Bernstein issue #7086 by @chernistry in #7101
- Bernstein issue #7092 by @chernistry in #7102
- Bernstein issue #7089 by @chernistry in #7103
- Bernstein issue #7095 by @chernistry in #7105
- Bernstein issue #7097 by @chernistry in #7104
- test: cover deterministic runtime and evidence boundaries by @chernistry in #7109
- docs: curate unreleased changes from merged pull requests by @chernistry in #7107
- docs: correct and complete Python docstrings by @chernistry in #7108
- Bernstein issue #7100 by @chernistry in #7110
- docs: correct CLI commands in guides and examples by @chernistry in #7111
- fix(cli): clarify help, diagnostics and log messages by @chernistry in #7112
- refactor: remove unused private helpers by @chernistry in #7125
- docs: put installation before the workflow example by @chernistry in #7126
- test: consolidate duplicate cases and parametrize lookup tests by @chernistry in #7129
- fix(cli): report gate and review verification failures on stderr by @chernistry in #7128
- fix(packaging): declare direct runtime imports and defer test helpers by @chernistry in #7127
- fix(cli): report unreadable input and missing workspaces on stderr by @chernistry in #7134
- fix(cli): send ledger and memory verification diagnostics to stderr by @chernistry in #7135
- fix: report fallback errors in local diagnostics by @chernistry in #7139
- fix(workflow): add a deadline to manifest predicates by @chernistry in #7141
- Bernstein issue #7137 by @chernistry in #7146
- fix(cli): send run bootstrap and task server errors to stderr by @chernistry in #7136
- fix: bound SDK and research reader requests by @chernistry in #7145
- Bind web approval decisions to queued requests by @chernistry in #7073
- Bernstein issue #7142 by @chernistry in #7147
- fix: add deadlines to finite maintenance commands by @chernistry in #7144
- Bernstein issue #7143 by @chernistry in #7151
- docs: repair the contributor setup and checks by @chernistry in #7149
- fix(schedule): resume cron jobs on the next leap day by @chernistry in #7152
- docs: align public API docstrings with implementation by @chernistry in #7148
- fix(git): preserve filenames in merge preview conflicts by @chernistry in #7154
- Keep command payloads out of application logs by @chernistry in #7157
- Cache locked CI dependencies by @chernistry in #7155
- Keep provider payloads out of application logs by @chernistry in #7158
- Verify receipt files in Governance by @chernistry in #7078
- Limit PR policy token to read access by @chernistry in #7156
- Omit transport and test payloads from logs by @chernistry in #7162
- Bernstein issue #7153 by @chernistry in #7161
- fix(schedule): isolate concurrent schedule file writes by @chernistry in #7163
- Remove unconditional orchestration diagnostics by @chernistry in #7164
- fix(config): report seed files with invalid UTF-8 by @chernistry in #7171
- Use lazy formatting for audited log calls by @chernistry in #7172
- Align runtime diagnostic levels with outcomes by @chernistry in #7170
- test: make queue and deadline checks deterministic by @chernistry in #7173
- Redact messages on Bernstein log handlers by @chernistry in #7160
- Bernstein issue #7166 by @chernistry in #7181
- Bernstein issue #7168 by @chernistry in #7184
- test: fix the date in metrics JSONL checks by @chernistry in #7175
- Bernstein issue #7179 by @chernistry in #7186
- docs: align configuration examples with the runtime by @chernistry in #7180
- test: synchronize wake and heartbeat assertions by @chernistry in #7176
- Bernstein issue #7188 by @chernistry in #7194
- test: verify regression guarantees without wall clocks by @chernistry in #7189
- test: scope audit and WAL property temp directories by @chernistry in #7183
- Bernstein issue #7165 by @chernistry in #7185
- test: make CLI status polling deterministic by @chernistry in #7174
- Bernstein issue #7192 by @chernistry in #7197
- test: control import and bootstrap fixtures by @chernistry in #7182
- test: isolate lineage examples and seed run-name samples by @chernistry in #7195
- Bernstein issue #7187 by @chernistry in #7196
- Keep keyboard focus in the approval queue by @chernistry in #7198
- Bernstein issue #7200 by @chernistry in #7208
- Reserve listeners in the two-node cluster fixture by @chernistry in #7199
- Bernstein issue #7205 by @chernistry in #7215
- Bernstein issue #7203 by @chernistry in #7216
- Bernstein issue #7150 by @chernistry in #7209
- test(contract): add contracts for cline, autohand, openhands, cody, composio and ralphex by @vaibhav8a in #6593
- Bernstein issue #7207 by @chernistry in #7214
- Bernstein issue #7138 by @chernistry in #7140
- Keep highlighted web text readable in both themes by @chernistry in #7217
- fix(tasks): keep dependents blocked on remaining failures by @chernistry in #7219
- fix(schedule): preserve the catch-up cursor across ticks by @chernistry in #7218
- fix(recap): read an explicitly supplied local archive by @chernistry in #7229
- fix(worktrees): roll back a failed checkout by @chernistry in #7227
- Bernstein issue #7237 by @chernistry in #7241
- Bernstein issue #7236 by @chernistry in #7242
- Bernstein issue #7234 by @chernistry in #7243
- Bernstein issue #7232 by @chernistry in #7244
- Bernstein issue #7231 by @chernistry in #7245
- Bernstein issue #7230 by @chernistry in #7246
- Bernstein issue #7239 by @chernistry in #7262
- Bernstein issue #7238 by @chernistry in #7266
- Bernstein issue #7252 by @chernistry in #7268
- Bernstein issue #7249 by @chernistry in #7269
- Bernstein issue #7248 by @chernistry in #7270
- Bernstein issue #7251 by @chernistry in #7271
- Bernstein issue #7254 by @chernistry in #7272
- fix(schedule): serialize updates to one schedule record by @chernistry in #7253
- fix(recap): preserve complete file paths in diff statistics by @chernistry in #7273
- Bernstein issue #7255 by @chernistry in #7274
- Bernstein issue #7260 by @chernistry in #7278
- Bernstein issue #7265 by @chernistry in #7279
- Bernstein issue #7263 by @chernistry in #7280
- Bernstein issue #7261 by @chernistry in #7281
- Bernstein issue #7259 by @chernistry in #7283
- Bernstein issue #7264 by @chernistry in #7277
- Bernstein issue #7258 by @chernistry in #7282
- Bernstein issue #7257 by @chernistry in #7284
- docs: correct verified API and CLI examples by @chernistry in #7297
- Bernstein issue #7286 by @chernistry in #7327
- Bernstein issue #7240 by @chernistry in #7328
- Bernstein issue #7291 by @chernistry in #7329
- Bernstein issue #7288 by @chernistry in #7326
- chore: reconcile repository templates and tooling by @chernistry in #7325
- Create tasks from the web task list by @chernistry in #7226
- Bernstein issue #7300 by @chernistry in #7332
- Bernstein issue #7293 by @chernistry in #7333
- Bernstein issue #7292 by @chernistry in #7336
- Bernstein issue #7295 by @chernistry in #7335
- Bernstein issue #7298 by @chernistry in #7331
- Refresh web UI screenshots for the documentation by @chernistry in #7337
- Bernstein issue #7324 by @chernistry in #7339
- Bernstein issue #7306 by @chernistry in #7340
- Bernstein issue #7302 by @chernistry in #7341
- Bernstein issue #7303 by @chernistry in #7342
- Bernstein issue #7304 by @chernistry in #7344
- Bernstein issue #7338 by @chernistry in #7346
- Bernstein issue #7345 by @chernistry in #7348
- Bernstein issue #7349 by @chernistry in #7376
- Bernstein issue #7352 by @chernistry in #7378
- Bernstein issue #7363 by @chernistry in #7379
- Bernstein issue #7353 by @chernistry in #7377
- Bernstein issue #7368 by @chernistry in #7381
- Bernstein issue #7369 by @chernistry in #7382
- Bernstein issue #7370 by @chernistry in #7384
- Bernstein issue #7380 by @chernistry in #7385
- Bernstein issue #7364 by @chernistry in #7387
- Bernstein issue #7354 by @chernistry in #7389
- Bernstein issue #7383 by @chernistry in #7388
- Bernstein issue #7392 by @chernistry in #7393
- Bernstein issue #7358 by @chernistry in #7394
- Bernstein issue #7375 by @chernistry in #7395
- Bernstein issue #7372 by @chernistry in #7396
- Bernstein issue #7365 by @chernistry in #7397
- Bernstein issue #7359 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7399
- Bernstein issue #7398 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7401
- Bernstein issue #7360 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7404
- Bernstein issue #7400 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7403
- Bernstein issue #7405 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7407
- Bernstein issue #7361 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7409
- Bernstein issue #7406 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7410
- Bernstein issue #7408 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7411
- Bernstein issue #7412 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7416
- Bernstein issue #7362 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7418
- Bernstein issue #7413 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7419
- Bernstein issue #7415 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7420
- Bernstein issue #7417 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7423
- docs: README banner by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7330
- Bernstein issue #7421 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7424
- Bernstein issue #7425 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7429
- Bernstein issue #7426 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7430
- Bernstein issue #7427 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7431
- Bernstein issue #7432 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7436
- Bernstein issue #7433 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7438
- Bernstein issue #7437 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7440
- Bernstein issue #7439 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7441
- Bernstein issue #7435 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7445
- Bernstein issue #7443 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7446
- Bernstein issue #7444 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7447
- Bernstein issue #7448 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7452
- Bernstein issue #7450 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7454
- Bernstein issue #7455 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7457
- Bernstein issue #7453 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7461
- Bernstein issue #7451 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7458
- Bernstein issue #7456 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7462
- Bernstein issue #7464 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7468
- Bernstein issue #7460 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7467
- Bernstein issue #7469 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7473
- Bernstein issue #7470 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7476
- Bernstein issue #7472 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7478
- Bernstein issue #7474 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7479
- Bernstein issue #7466 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7480
- Bernstein issue #7463 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7482
- Bernstein issue #7465 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7483
- Bernstein issue #7481 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7486
- Bernstein issue #7485 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7494
- Bernstein issue #7490 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7496
- Bernstein issue #7495 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7497
- Bernstein issue #7499 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7503
- Bernstein issue #7504 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7505
- Bernstein issue #7501 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7507
- Bernstein issue #7502 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7508
- Bernstein issue #7510 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7512
- Bernstein issue #7506 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7513
- Bernstein issue #7514 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7518
- Bernstein issue #7511 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7519
- Bernstein issue #7515 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7522
- Bernstein issue #7520 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7525
- Bernstein issue #7524 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7528
- Bernstein issue #7517 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7529
- Bernstein issue #7527 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7532
- Bernstein issue #7526 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7535
- Bernstein issue #7531 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7536
- Bernstein issue #7530 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7537
- Bernstein issue #7538 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7541
- Bernstein issue #7533 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7542
- Bernstein issue #7539 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7544
- Bernstein issue #7534 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7545
- Bernstein issue #7543 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7549
- Bernstein issue #7547 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7551
- Bernstein issue #7540 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7550
- Bernstein issue #7548 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7553
- Bernstein issue #7552 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7557
- Bernstein issue #7546 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7558
- Bernstein issue #7555 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7561
- Bernstein issue #7554 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7559
- Bernstein issue #7556 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7562
- Bernstein issue #7560 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7564
- Bernstein issue #7565 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7569
- Bernstein issue #7563 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7570
- Bernstein issue #7566 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7572
- Bernstein issue #7567 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7574
- Bernstein issue #7568 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7575
- Bernstein issue #7571 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7577
- Bernstein issue #7578 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7581
- Bernstein issue #7576 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7582
- Bernstein issue #7573 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7583
- Bernstein issue #7580 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7584
- Bernstein issue #7579 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7587
- Bernstein issue #7585 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7591
- Bernstein issue #7589 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7593
- Bernstein issue #7296 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7334
- Bernstein issue #7592 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7596
- Bernstein issue #7588 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7598
- Bernstein issue #7590 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7597
- Bernstein issue #7422 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7428
- Bernstein issue #7487 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7493
- Bernstein issue #7471 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7489
- Bernstein issue #7599 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7603
- Bernstein issue #7594 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7604
- fix(routes): keep remote branch fetch logs on one line by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7068
- Bernstein issue #7601 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7608
- Bernstein issue #7595 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7609
- Bernstein issue #7600 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7610
- Bernstein issue #7606 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7612
- Bernstein issue #7605 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7613
- Bernstein issue #7611 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7615
- Bernstein issue #7614 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7617
- Bernstein issue #7616 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7618
- fix(ledger): resume after an interrupted append by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7276
- fix(adapters): keep multiline standing limits from retrying by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7275
- feat(adapters): run OpenClaw headlessly with isolated state by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/6268
- Bernstein issue #7620 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7623
- Bernstein issue #7619 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7625
- fix(live): show remote worker nodes in the AGENTS panel by @chernistry in #6753
- Bernstein issue #7622 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7628
- Bernstein issue #7621 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7629
- Bernstein issue #7626 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7630
- fix(typing): preserve the caching adapter vendor interface by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7072
- test: check metric files by prefix by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7220
- test: cover the exact tick budget boundary by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7221
- test: isolate the zero-timeout capacity wait by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7222
- fix(typing): annotate the telemetry span iterator by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7123
- test(metrics): make queue depth timestamps deterministic by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7124
- fix(plans): reference plan decision API routes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7075
- docs: align README name section with governance-layer wording by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7635
- Bernstein issue #7627 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7632
- Bernstein issue #7631 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7634
- chore(deps): update github/codeql-action digest to 33d7b85 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7002
- Bernstein issue #7373 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7390
- Bernstein issue #7484 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7491
- Bernstein issue #7602 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7607
- fix(adapters): harden Claude commands without an MCP payload by @tettehbenjamin651-hash in https://github.com/sipyourdrink-ltd/bernstein/pull/6804
- docs(tests): refresh the context file against its subtree (#6939) by @vaibhav8a in https://github.com/sipyourdrink-ltd/bernstein/pull/6972
- Bernstein issue #7498 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7500
- ci: reduce runner contention in the merge queue by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7210
- Bernstein issue #7371 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7386
- docs: distinguish managed plan IDs from run IDs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7223
- docs: clarify the run lookup search scope by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7224
- docs: clarify Plane ticket IDs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7225
- docs: define batch and pass in the receipts guide by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7228
- docs: align volunteer examples with the CLI by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7074
- docs: document offline benchmark verification by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7076
- fix(identity): contain persisted identity record paths by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7070
- Apply trigger rules to generic webhook task creation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7021
- fix(admission): use the registered key for pinned adapters by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7071
- Bernstein issue #7477 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7488
- Bernstein issue #7475 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7492
- docs: regenerate adapter last-green table from canary receipts by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7077
- fix(adapters): stream Pi output in non-interactive mode by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7645
- fix(templates): align adapter scaffold with CLIAdapter by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7644
- fix(ci): require landed merge groups before release dispatch by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7652
- docs: clarify cleanup force and branch removal by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7660
- fix: retain unstable properties in render comparisons by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7661
- chore(deps): update dependency click to v8.5.1 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7646
- Preserve declared gate scopes across result handling by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7662
- docs: clarify product and workflow boundaries by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7666
- Allow custody signers for lineage detached JWS by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7664
- fix: wait for coupled trufflehog dependency updates by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7669
- Qualify the schedule catch-up list annotation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7668
- chore(deps): react 19.3.1 with compatibility fixes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7671
- chore(deps): kubernetes 37.0.1 with compatibility fixes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7675
- feat: record idempotent reconcile apply and partial resume by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7673
- fix(tasks): preserve unresolved retry inputs (#6605) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7674
- chore(deps): cosign 3.1.4 with compatibility fixes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7678
- chore(deps): react-query 5.105.3 with compatibility fixes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7684
- feat(eval): report recorded intent calibration by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7691
- feat(replay): compare supplied conduct activity scopes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7695
- fix(security): harden MCP definition observation (#6984) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7693
- Add explicit observed remote status and cancellation results by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7700
- feat(browser): add optional computed-style render probe by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7694
- fix(security): harden default adapter policy (#7019) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7686
- Select one hash profile for new run evidence by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7692
- feat: composition-only Agent Manifest export with operator declarations by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7697
- fix(security): harden doctor lane evidence (#6572) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7706
- Preserve task routing hints through persisted records by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7698
- chore(deps): mcp 2.3.0 with compatibility fixes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7707
- fix(security): harden MCP control mapping (#6217) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7708
- chore(deps): update actions/checkout action to v7 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7053
- fix(security): persist task claim admission control (#6987) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7712
- Synchronize process and cleanup fixtures by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7201
- docs(memory): document CrossTaskKB publish return and subscribe itera… by @Vivan217 in https://github.com/sipyourdrink-ltd/bernstein/pull/7347
- docs: clarify token growth sampling and warning reset by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7725
- docs: explain readback after an uncertain task creation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7726
- docs: handle approvals and terminal task polling outcomes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7727
- docs: describe the recorded run replay JSON envelope by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7728
- docs: describe the current task list readback limitation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7729
- docs: correct the web interface route inventory by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7730
- docs: keep the bundled GUI on the mock playground server by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7731
- docs: explain task search prefixes and status chips by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7732
- docs: align the task drawer action summary by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7733
- docs: distinguish plan steps workflow nodes and scenario templates by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7734
- docs: explain selected protocol patches and capability sets by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7735
- docs: explain adaptive parallelism state after restart by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7736
- docs: correct task action endpoint references by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7737
- docs: document notification sink lifecycle by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7738
- docs: distinguish catalog parsing defaults by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7739
- docs: clarify fast resume snapshot scope by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7740
- docs: distinguish ledger and replay journal roles by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7741
- docs: use recorded agent states for idle inspection by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7742
- docs: map tuning guidance to status report fields by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7743
- docs: explain recap data sources and local modes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7744
- docs: explain recipe preview inputs and output by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7745
- docs: clarify simulation formats and budget breach output by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7746
- docs: clarify comparison mode selection and output by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7747
- docs: explain diary project selection and inspection errors by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7748
- docs: distinguish run state from worker sessions by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7749
- docs: clarify tracker task normalization by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7750
- docs: distinguish scenario templates from emitted tasks by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7751
- docs: distinguish record identity from display number by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7752
- docs: link proposal review to rollback records by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7755
- docs: qualify worktree merge isolation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7756
- docs: explain browser dashboard workspace selection by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7757
- docs: qualify supported agent interchangeability by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7758
- docs: describe server driven editor budget notices by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7759
- docs: align editor setup terminology by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7760
- docs: name the deterministic scheduling component by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7761
- docs: distinguish test file and case selection by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7762
- docs: explain trace index rebuild recovery limits by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7763
- docs: link worktree inspection to parallel run graphs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7754
- docs: clarify integration host and tool server roles by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7753
- fix(identity): omit retired signing file paths from logs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7069
- chore(ci): update create-pull-request to v8.1.2 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7654
- docs: replace local machine paths with neutral paths by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7771
- ci: run CI on release/** branches by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7769
- docs(governance): temporary merge rule until 23 November for v4.0.0 stabilisation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7767
- test(tui): restore process streams after concurrent dashboard polls by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7721
- docs: link the bernstein-addons plugins by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7772
- docs: correct the bernstein-addons plugin descriptions by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7799
- chore(deps-dev): bump diff-cover from 10.6.0 to 11.0.1 by @dependabot[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7719
- chore(deps-dev): bump pypdf from 6.21.4 to 7.0.0 by @dependabot[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7717
- chore(deps): update dependency lucide-react to v1.71.0 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7711
- chore(deps-dev): bump types-pyyaml from 6.0.12.20260906 to 6.0.12.20261103 by @dependabot[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7718
- docs(engagements): the playbook README shows a command that does not exist by @vaibhav8a in https://github.com/sipyourdrink-ltd/bernstein/pull/7062
- fix(ci): collect coverage after the full main test matrix by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7802
- docs(community): narrow the decisions sentence in the recognition issue by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7804
- docs(skills): explain in-memory load failure reporting by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7808
- docs(git): explain shared snapshot metadata and fallback records by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7810
- fix(volunteer): cap JSON responses before decoding by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7809
- docs(community): remove the usage paragraph from the recognition issue by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7811
- Derive a structural plan graph from recorded events by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7680
- docs: remove stale awesome-python listing claim by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7815
- fix(cli): print the selectable adapter count and align documentation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7816
- fix(worker): retire the child PID before joining the log monitor by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7807
- chore(deps): update dependency vite to v8.3.8 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7710
- chore(deps): update dependency markdown to v3.11.3 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7679
- chore(deps): update dependency mkdocs-material to v9.8.1 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7687
- feat(cluster): add consistent hash ring primitive by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7648
- chore(deps): update dependency backrefs to v8.2.1 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7822
- Fix unconfirmed task lookup smoke synchronization by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7824
- docs(agy): state consumer-plan restrictions by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7818
- fix(devcontainer): default to the vscode user by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7806
- test(tui): make dashboard poll ownership assertion deterministic by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7828
- chore(deps): update dependency @vitejs/plugin-react to v6.3.0 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7826
- chore(deps): update dependency @types/vscode to v1.146.0 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7825
- chore(deps): update grafana/grafana docker tag to v13.3.1 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7823
- chore(deps): bump the python-minor-and-patch group across 1 directory with 25 updates by @dependabot[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7821
- fix(cli): select the installed prerelease notes offline by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7832
- build(devcontainer): pin the base image by digest by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7833
- test(workflows): characterize dispatch crash recovery by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7659
- feat(replay): record unreadable runs in conduct aggregation by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7704
- Route CMP observations through the shared audit checks by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7665
- Clarify README dashboard and MCP interfaces by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7688
- Preserve dispatch tiers through retries and explicit reopen by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7701
- docs: regenerate adapter last-green table from canary receipts by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7713
- docs(storage): document ArtifactSink read and write contract by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7702
- fix(security): harden contribution quorum (#7020) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7683
- fix(security): harden independent bot reviews (#6761) by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7709
- chore(deps): update otel/opentelemetry-collector-contrib docker tag to v0.165.0 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7831
- docs(workflows): preserve goal substitution on resume by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7840
- docs: preserve the original goal in README workflow resume examples by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7838
- chore(deps): update dependency ovsx to v1.4.0 by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7830
- docs: describe the conductor metaphor as orchestration by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7839
- chore(deps): update python:3.13-slim docker digest to e48f59c by @renovate[bot] in https://github.com/sipyourdrink-ltd/bernstein/pull/7667
- feat(compliance): surface declaring suites and latest measurements in compliance controls by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7676
- test(workflows): characterize completion authority by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7649
- feat(cluster): evaluate worker health for scheduling by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7656
- docs(agents-md): add review guidelines for Codex code review by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7852
- Show project readiness and test the first run through recap by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7846
- Reword the leaderboard footer by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7835
- chore(ci): update typos action to v1.52.0 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7650
- docs: correct CONFIG, MCP server and plugin SDK pages by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7861
- test: cover git snapshot and SDK2 legacy task paths by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7860
- examples: add four runnable offline demos by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7862
- chore(ci): update actionlint action to v1.83.0 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7651
- docs: describe Bernstein as an open-source governance layer by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7863
- Count increases in reported cumulative run cost by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7864
- Record failed detached tasks when the SSH runner returns a nonzero exit by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7866
- docs: describe the Tasks diff drawer in the homepage screenshot by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7877
- docs(cli): correct template command arguments by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7876
- Document how to use bernstein explain by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7880
- Clarify that runs report includes active runs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7879
- Clarify the file lookup in history help by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7878
- Document and test the main report command by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7881
- Document and test the managed plan state filter by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7884
- Test the cache list result limit by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7885
- Disable the shared uv cache for the pipx wheel smoke job by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7896
- Document the upgrade from Bernstein 3.24 to 4.0 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7895
- Test the flags for diff comparisons by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7886
- Omit credential selectors from OpenAI Agents spawn logs by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7897
- Test recap with an explicit project directory by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7887
- Test the JSON and server URL options for cluster status by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7888
- Test the CLI flags for API and dependency impact by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7891
- Clarify and test simulation output options by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7889
- Test the comparison output directory option by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7894
- Align live help with the display modes by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7893
- Document the JSON configuration path output by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7892
- Update the workflow lint action to v1.84.1 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7905
- Test scaffold overwrite behavior by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7890
- Pin setup-uv to v10.3.3 and clarify bootstrap retries by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7908
- Pin create-pull-request to v8.1.3 in five workflows by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7907
- chore(release): v4.0.0 by @chernistry in https://github.com/sipyourdrink-ltd/bernstein/pull/7843
New Contributors
- @aarav570 made their first contribution in #6803
- @tettehbenjamin651-hash made their first contribution in https://github.com/sipyourdrink-ltd/bernstein/pull/6804
- @Vivan217 made their first contribution in https://github.com/sipyourdrink-ltd/bernstein/pull/7347
Full Changelog: v3.24.1...v4.0.0