Repository navigation
feat(cli): move workflow consumers to Relayflows v2 - #2104
Conversation
Remove crates/broker, crates/relay-pty, the root Cargo workspace and Cross.toml, and the broker-<platform>, harness-driver, harnesses, fleet and evals packages, along with their workspace, lockfile, CI and publish references. publish.yml keeps build-node; create-release and summary no longer wait on broker jobs, and the version step leaves the frozen harness-driver and broker-<platform> pins at 17.2.1. The broker-only CLI commands stay registered and print a removal notice pointing at `relay init` and `relay init --hosted`. relayflows, cloud-worker and `node workflow run` keep @relayflows/core 1.1.10, which resolves harness-driver and broker-<platform> 17.2.1 from npm. Delete fleet-daytona.mjs and the cleanroom qualification consumer workflow, move install-paths into the CLI, and reword the vendored orchestrating-agent-relay skill and the durable task provider spec. crates/relay-node is untouched. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the
|
Codex Review Summary: ❌ FailedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
[Medium impact] The PR changes how local and Cloud worker workflow runs start. Stop the standalone workflow child before merging; worker shutdown currently leaves it running during cleanup.
|
There was a problem hiding this comment.
All reported issues were addressed
Tip: Get a Slack DM when cubic finishes reviewing your PRs. Connect Slack
View guided diff | Re-trigger cubic
e7ec4d1 to
43a614d
Compare
There was a problem hiding this comment.
All reported issues were addressed
Note: This PR contains a large number of files. cubic selects up to 200 of the highest-priority eligible files for this review, so some files may not have been reviewed.
View guided diff | Re-trigger cubic
There was a problem hiding this comment.
All reported issues were addressed
Note: This PR contains a large number of files. cubic selects up to 200 of the highest-priority eligible files for this review, so some files may not have been reviewed.
View guided diff | Re-trigger cubic
# Conflicts: # package-lock.json
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
| const agentRelayCli = resolveAgentRelayCliChildCommand({ | ||
| execPath: process.execPath, | ||
| cliEntrypoint: deps.resolveAgentRelayCliEntrypoint(), | ||
| args: buildRelayflowsRunArgs(workflowPath, payload.fileType), | ||
| cwd: deps.cwd(), | ||
| }); | ||
| const result = await runChild({ | ||
| command: process.execPath, | ||
| args: relayflowsArgs(relayflowsCli, workflowPath, payload), | ||
| command: agentRelayCli.command, | ||
| args: agentRelayCli.args, |
There was a problem hiding this comment.
Workflow survives worker shutdown
Stopping a standalone Cloud worker now kills the agent-relay flows run wrapper, not the workflow. The standalone product loader starts a separate Node child, but neither the loader nor CLI startup forwards SIGTERM to it. The worker then removes the run directory and login files while that child can keep running.
Forward cancellation to the actual runner and wait for it to stop before cleanup.
Knowledge Base Used: CLI command surface and runtime utilities
Prompt To Fix With AI
This is a comment left during a code review.
Path: packages/cli/src/cli/commands/cloud-worker.ts
Line: 253-261
Comment:
**Workflow survives worker shutdown**
Stopping a standalone Cloud worker now kills the `agent-relay flows run` wrapper, not the workflow. The standalone product loader starts a separate Node child, but neither the loader nor CLI startup forwards `SIGTERM` to it. The worker then removes the run directory and login files while that child can keep running.
Forward cancellation to the actual runner and wait for it to stop before cleanup.
**Knowledge Base Used:** [CLI command surface and runtime utilities](https://app.greptile.com/-/switch-tenant?tenant=agent-relay&next=%2Fcustom-context%2Fknowledge-base%2Fagentworkforce%2Frelay%2F-%2Fdocs%2Fcli-command-surface.md)
---
For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.
Summary
@relayflows/cli/@relayflows/corerunneragent-relay node workflow runand fresh Cloud worker assignments through the existing v2agent-relay flows runproduct surface/$bunfsscript, and use Node for standalone JavaScript workflow/monitor scriptsStack / release boundary
chore/retire-broker-v2; this remains separate from the broker-retirement change@relayflows/sdkstays at the currently published exact pin (2.0.65) while flows#707 is a draft; update the pin to its released version before this PR leaves draftVerification
npm cisucceeds and installs neithernode_modules/@relayflows/clinornode_modules/@relayflows/corenpm run typecheckpassesnpm run build:clipassesnpm pack --dry-runpassesgit diff --checkpass on changed filesCoordination
startFrom/previousRunId, and keeps resume state in its hosted runner