Repository navigation
Promote expbkmain: reconnects no longer wait on editor and SSH discovery (#248), plus #240 and #244 - #249
Merged
Conversation
…ale owner All six deploy scripts pinned the GitHub API to beknown-work/bkt3code. The repository has since been transferred to iamtushar324, so that path answers with a 301 and nothing else. auto-deploy.sh read it with a curl that did not follow redirects, so jq got the redirect body and failed with "Cannot iterate over null" on every timer tick; bkt3 sat on old code after a green build until the artifact was installed by hand. The redirect-following flag has since been added, but every script still depends on a redirect GitHub only keeps until someone claims the old name. Each script now derives the slug from its checkout's origin remote and fails with a clear message if that is not an owner/name pair, so the next transfer needs no change here. Model: Claude Opus 5 via Claude Code in T3 Code. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
fix(deploy): deploy scripts read the repository from origin, not a stale owner
Every x-bf-vk integration was forced to bk-toolhub, so nobody could try toolyard, the MCP gateway replacing Bifrost, without a code change. The integration may now point at an allowlisted Bifrost-compatible gateway (Bifrost, the default, or toolyard.dev.beknown.live/mcp), picked per person in Settings; any other URL still rewrites to Bifrost, matching is exact after normalisation, and only the canonical URL is stored. The integration id, auth mode, credential storage and the `bifrost` tool namespace are unchanged, so skills keep working. Done by Claude Opus 5.5 in Claude Code (T3 Code session). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
feat(mcp): let a user route their Bifrost integration through toolyard
Every websocket connect built the server config by scanning PATH for editors and probing sshd/Tailscale inline, each under a 5 s timeout. On a busy server those steps timed out, the config took 12 s or more, and clients (15 s setup budget) gave up and retried; a timed-out scan was never cached, so every retry paid again. bkt3 showed "disconnected" for ~27 s. Serve each discovery from its last success and refresh it in the background after a minute. Only the first call after startup waits, and its scan runs in the layer scope, so a caller that gives up no longer cancels it. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…nput Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…covery fix(server): reconnects no longer wait on editor and SSH discovery
This branch was successfully deployed
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.
Promotes
expbkmain(9d23bf0be) tobkmain. All three changes are already running on expbkt3.Verified on expbkt3 (
9d23bf0be, deployed 14:27 UTC)server.getConfigtakes 1–2 ms. Before the fix it took 16–17 ms, because the SSH/Tailscale probe ran on every connect.getConfiganswered in 2 ms (14:29:57.887–.889). The editor refresh started after that, at .900–.970. A second connect during the refresh answered in 1 ms and did not start another scan. Before the fix, the same connect ran the scan inline: 21 PATH checks insidegetConfig, 45 ms idle and up to 5 s under bkt3's load.Merging this
Merging deploys bkt3, which restarts
t3-bkmain.serviceand interrupts every live session on it.🤖 Generated with Claude Code — Claude Opus 5.5 in T3 Code