The exit callback enqueued the exited frame before deleting c.terms[id], so a client reopening on receipt of exited could hit the open guard while the entry was still set and have its open dropped. Delete first so the cleared entry is visible by the time the client sees exited. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> |
||
|---|---|---|
| .github/workflows | ||
| backend | ||
| docs | ||
| frontend | ||
| .gitignore | ||
| README.md | ||
README.md
agent-orchestrator
Rewrite of the agent-orchestrator: a long-running Go backend daemon (backend/)
paired with an Electron + TypeScript frontend (frontend/).
See docs/ for architecture and status — start with the
Lifecycle Manager + Session Manager lane in docs/architecture.md.
Backend daemon
The Go binary in backend/ is the HTTP daemon — a loopback-only
sidecar the Electron supervisor will spawn (Phase 1c). Phase 1a landed the
skeleton: chi router, middleware stack (recoverer → request-id → logger →
real-ip), /healthz + /readyz, atomic running.json PID/port handshake,
graceful shutdown on SIGINT/SIGTERM.
Run
cd backend
go run . # binds 127.0.0.1:3001 with all defaults
AO_PORT=3019 go run . # override per invocation
Health check:
curl localhost:3001/healthz # {"status":"ok"}
curl localhost:3001/readyz # {"status":"ready"}
Configuration (env only)
The bind host is always 127.0.0.1: the daemon is a loopback-only sidecar
and binding any other interface would be a security regression, so the host
is intentionally not env-configurable.
| Var | Default | Purpose |
|---|---|---|
AO_PORT |
3001 |
bind port; fails fast if taken |
AO_REQUEST_TIMEOUT |
60s |
per-request timeout (Go duration) |
AO_SHUTDOWN_TIMEOUT |
10s |
graceful-shutdown hard cap |
AO_RUN_FILE |
<UserConfigDir>/agent-orchestrator/running.json |
PID + port handshake path |
Test
cd backend
gofmt -l . && go build ./... && go vet ./... && go test -race ./...