Harden factory memory, relay recovery and integration checks - #14
Merged
Conversation
Adapt block/buzz block#6950 memory guidance and fixture to the existing benchmark runtime. Backport block#7010 receive-boundary verification and block#7325 targeted overflow recovery, preserving fork lifecycle protections. Signed-off-by: orchestrator <neo870123@gmail.com>
Adapt MeshLLM migration from block/buzz block#6189 and model translation from block#5289. Update h2 and webbrowser in both dependency graphs, exercise mesh-enabled desktop tests, and use configured fork namespaces for Docker caches and provenance. Signed-off-by: orchestrator <neo870123@gmail.com>
Signed-off-by: orchestrator <neo870123@gmail.com>
Theremoteaidoc
marked this pull request as ready for review
September 8, 2026 23:08
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Relay events could reach agent queues without signature verification, overflow recovery could replay unrelated subscriptions, and agents lacked accurate cold-memory retrieval guidance. This change verifies incoming events before routing or state mutation, retries only affected subscriptions when capacity returns, and documents the actual memory CLI with an isolated retrieval benchmark.
Backports/adaptations: block/buzz block#6950, block#7010 and block#7325. The required MeshLLM/Nostr migration follows block#6189 with model translation from block#5289; stored
autoand blank models become wire modelmesh, while named models are preserved. Both lockfiles update h2/webbrowser and remove the abandoned nostr-relay-pool dependency. Mesh remains opt-in. Existing fork heartbeat, owner-interrupt, WIP and turn-limit behavior is retained.The Docker workflow now routes gateway metadata, caches, digests and provenance through the fork's configured registry, matching the relay override. Previously builds succeeded and then failed exporting cache to block's registry. Full mesh-enabled desktop library tests now run in CI. Environment handling was extracted into the existing module to keep the runtime file below its prior size without weakening the size gate.
Validation on ci1 as buzzagent: 767 ACP tests, 2,331 desktop mesh library tests (14 pre-existing ignored), final mapping helper tests in both feature configurations, root relay all-targets compile, formatting, clippy and pinned dependency policy checks passed. Python benchmark tests/lint and both gateway builds also pass on GitHub. Independent verification confirmed clean final source SHA before and after testing. All 27 expected GitHub PR checks passed before merge; the only skipped jobs were unchanged Web paths and image publication disabled for pull requests.
The final sprig candidate was built from
30dd024881ad55139eac7d8baeb397401ceb7d8b; binary SHA256 isea221c0ec06406b1709a06bedcca22acacae75f843a39f4ec34aac8645bbaf92. Installed on all seven core factory agents and the Hostinger Hermes bridge after merge, with exact process executable identity and new idle heartbeats verified. Protected rollback artifacts are retained; the retired Codex seat stays disabled. Shared factory AGENTS/system instructions are maintained separately in the ops repository and their installed hashes have been checked.Memory fixture tests do not establish live model retrieval success. Overflow replay remains best effort; a successful request write is not a delivery receipt. No product promotion or live test task was dispatched.