feat(fleet): qualify scoped dispatch with durable Buzz receipts - #19
Draft
mfethe1 wants to merge 2 commits into
Draft
feat(fleet): qualify scoped dispatch with durable Buzz receipts#19mfethe1 wants to merge 2 commits into
mfethe1 wants to merge 2 commits into
Conversation
Signed-off-by: Michael Feth <mfethe1@gmail.com>
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: trueThanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This was referenced Sep 7, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Buzz CML tasks currently have no qualified path into the fleet’s durable execution receipts. Add an on-demand adapter that verifies a signed plan and host-local scope, freezes outgoing transitions before publication, records an attempt before execution, and stores its result before submitting signed evidence for review. Stable attempt IDs and payload digests prevent duplicate delivery from re-executing completed work and refuse automatic replay of an unknown outcome.
This initial capability only reads an approved repository’s HEAD and indexed entry count. It does not run models, modify repositories, or approve completion. Windows admission is blocked until process-tree containment is qualified. POSIX process groups are only used for the fixed probe; this is not a sandbox or a server-side atomic lease for arbitrary agents.
Related issue
Related foundations: #10 and #15. Searched existing issues and PRs; no duplicate durable qualification adapter found.
Testing
At head
7fce31123409df3c2e32a176099d9b80473b79dd, all 16 adapter tests passed, including duplicate delivery, lost publication replies, scope tampering, unknown outcomes, and cancellation. Independent review reproduced the Git-filter execution risk and Windows containment gap, then verified the fixed probe and Windows admission denial.A native signed plan-to-review run passed. The scoped network candidate also completed Mack scheduler dispatch to Mack, Airy, and Rosie. Each host's receipt used the same task UUID as its signed CML task and human task record; the task comment records the attempt and signed result IDs. Replaying the finished envelope to each installed host preserved receipt timestamps, results, journals, and outgoing-event records. All three installed hosts identically rejected eight scope expansions. An independent reviewer checked Mack's live receipt, installed files, signed reduction, task, and channel memberships. CML remains
review, human tasks remainin_progress, and acceptance remains unverified. These network checks used separately staged operator wrappers; this PR adds the adapter, not a continuous scheduler or task-row subscription.Full
just cifailed at this exact head: the mobile suite reported 2,241 passed and eight failed agent-activity golden comparisons. The same eight tests fail at unchanged baseadeec6eb5c4ff425aa03d8a744a295da4429f67f; this PR does not change mobile files or expected images. The network probe exercised no model inference or provider fallback. Rosie used an index-only clone, and Airy used the temporary candidate checkout; physical-phone operation and Windows execution remain unqualified.