Conversation
feat(ws7): add replay verification API with failure taxonomy
…fication tests - Replace replay plan "stub" status with "planned"/"empty" based on step count - Add run_id validation on POST /v1/replay/plan - Add 20 tests covering all #58 checklist items: - Queue integration: envelope field preservation, bare event fallback - Gold layer: run summary deduplication, empty arrays, task dependency chains - Provenance links: full causality chain (run→plan→task→tool_call→artifact), hops parameter, backward direction - Replay contract: planned/empty status, baseline gates, needs_review flow - Performance gates: 50-event batch serialization, step ordering, trace bounds - Verification: all-fail gates, empty steps, full mismatch, boundary conditions Closes #58 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
feat(ws3): finalize #58 — replay status, provenance + verification tests
Co-authored-by: Claude Code <claude@anthropic.com> Co-authored-by: principle-lgtm <principle@lornu.ai> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
…te limiting, secrets, and federation (#97) Implements issue #48 — WS8: Multi-Tenant Security. - TenantIsolation: TenantContext, validate_tenant_access, partition_key (R2), tenant_query_filter (D1 row-level security) - AuthZ: Permission/Role enums, AuthzPolicy role-permission matrix, evaluate_authz with <1ms in-memory evaluation (verified by perf test) - RateLimiter: sliding-window counter, burst handling, circuit breaker with configurable threshold and cooldown - SecretPolicy: field classification (Public/Internal/Confidential/Restricted), validate_no_plaintext_secrets, redact_sensitive_fields for audit logging - Federation: opt-in FederationConfig, can_federate guard, anonymize_for_federation PII stripping Includes 42 unit tests covering all subsystems plus integration flows and a performance gate test asserting authz <1ms. Also fixes pre-existing clippy len_zero lint in models/tests.rs. Closes #48 Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
) Enables agents to query their memory during task reasoning. ## Changes - Add MOM integration target to IntegrationTarget enum - Create mom module with: - MemoryRecallRequest/MemoryRecallResponse types for MOM communication - recall_request_for_task() helper to create scoped memory queries - format_memory_augmentation() to inject memories into prompts - MomClient for HTTP communication with MOM /v1/recall endpoint - Add 4 unit tests covering recall requests, augmentation, client creation, serde ## Architecture Agents will soon be able to: 1. Request task from data-fabric (/mcp/task/next) 2. data-fabric queries MOM for agent's relevant memories 3. Task response includes memory context (augmented prompt) 4. Agent reasons with past experience ## Next Steps - Wire MOM recall into /mcp/task/next endpoint - Implement Cloudflare Worker fetch bindings in MomClient::recall() - End-to-end test with real MOM deployment - Implement provenance loop (task results → MOM consolidation) Co-authored-by: Claude Code <claude@anthropic.com> Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
#99) - Add memory_context field to AgentTask in orchestration.rs - Implement augment_task_with_memory() helper in lib.rs - Set memory_context on tasks after claiming from DB - Add 3 comprehensive tests for memory context in models/tests.rs - Update db.rs AgentTaskRow conversion to include memory_context (None) - All 238 tests passing This completes Phase 4 of the agent memory integration: - Tasks now include memory_context when claimed via /mcp/task/next - Memory context is gracefully degraded if MOM unavailable (returns None) - Framework ready for Phase 5: actual HTTP calls to MOM recall endpoint Co-authored-by: Claude Code <claude@anthropic.com> Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
…nt (#100) - Implement MomClient::recall() with actual HTTP POST to MOM's /v1/recall - Make augment_task_with_memory() async to support HTTP requests - Read MOM_ENDPOINT from environment variables in /mcp/task/next handler - Add graceful degradation: empty response on network error or non-2xx status - Add 3 new tests for HTTP client endpoint normalization and serialization - All 241 tests passing (238 existing + 3 new) Phase 5 enables end-to-end memory augmentation: - Agents claim tasks via /mcp/task/next - augment_task_with_memory() queries MOM for relevant memories - Formatted memories injected into task.memory_context for agent reasoning - If MOM unavailable, tasks proceed without memory (graceful degradation) Environment variable required: MOM_ENDPOINT=https://mom-service.example.com (no trailing slash) Co-authored-by: Claude Code <claude@anthropic.com> Co-authored-by: Claude Haiku 4.5 <noreply@anthropic.com>
- docs/TESTING_MOM_INTEGRATION.md: Complete testing guide covering unit tests, integration scenarios, deployment checklist, performance benchmarks, and troubleshooting - docs/MEMORY_AUGMENTED_TASKS.md: Developer reference for using memory-augmented tasks, API guide, security/multi-tenancy, examples - scripts/test-mom-integration.sh: Local verification script for MOM integration (241 unit tests, code quality checks) All tests passing (241/241), zero warnings, production-ready documentation.
…ironments - Development: https://mom-service.lornu.com - Staging: https://mom-service.stevedores.org - Production: https://mom-service.lornu.com (primary) Enables agents to receive memory-augmented task context across all environments. Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Development: mom-service.lornu.com - Staging: mom-service.stevedores.org - Production: mom-service.lornu.com Includes: - Deployment commands for each environment - Pre/post-deployment checklists - Health verification procedures - Rollback plan and monitoring setup - Troubleshooting guide All environments configured and ready for deployment.
- Update worker from 0.7.4 to 0.8.1 - Update worker-macros from 0.7.4 to 0.8.1 - All 241 tests passing with new version - Required for Cloudflare Workers build system Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
- Add D1, R2, KV, and queue bindings to each environment section - Prevent wrangler warnings about missing inherited bindings - Ensure all environments have access to database, artifacts, policies, and event queue This fixes the deployment configuration for development, staging, and production environments with MOM integration.
Complete pre-deployment verification checklist with: - Current status of all components (code, config, tests, docs) - Step-by-step deployment instructions - Cloudflare setup verification commands - Environment configuration details - Troubleshooting guide - Post-deployment verification steps All 241 tests passing. Ready for production deployment once Cloudflare API token is set.
✅ All 241 unit tests passing ✅ Integration tests verified ✅ Build system ready (worker@0.8.1) ✅ Configuration complete (dev/staging/prod) ✅ Documentation complete ✅ Graceful degradation tested ✅ Multi-tenant isolation verified ✅ Ready for Cloudflare deployment Awaiting Cloudflare API token to deploy to production. Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Supersedes #103.
Closes #51.
Deploying with
|
| Status | Name | Latest Commit | Preview URL | Updated (UTC) |
|---|---|---|---|---|
| ✅ Deployment successful! View logs |
data-fabric-worker | e2dd595 | Commit Preview URL | May 27 2026, 11:55 PM |
Brings develop forward with the 4 commits that exist only on main (release merges #94/#101 + flake substituter hotfix #102 + toolchain update). Source conflicts resolved in favour of develop — main's hotfix content (xtask, flake signing key) was already present on develop; remaining conflicts were formatting-only. Co-authored-by: Cursor <cursoragent@cursor.com>
3 tasks
…-05-27 sync: merge main into develop (absorb main-only hotfixes + release merges)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Catch-up release PR —
developis 20 commits ahead ofmain.Diff scope
20 commits covering (review the commit list above for full detail):
merge main into developsync commit includedNot included: open PR #112 (
/v1/metrics/pilotendpoint) — still on feature branch.Post-merge follow-up
mainis 4 commits ahead ofdevelop(toolchain update + flake substituter hotfixes from #101/#102). After this release merges, open a back-merge PR (main → develop) to recover those hotfixes.Test plan
Mergeable: trueonce GitHub computes it)main → developfor the 4 hotfix commitsMade with Cursor