Your agent ran. It never saw your repo.
Workspace fallback fix + unofficial Docker self-host stack — attach the project folder first.
Hey {{firstName}},
Last week was secrets and quotas. This week is the quieter failure mode: the run finishes, the log still says workspace unavailable, and the agent never opened your tree.
Short issue. High leverage if onboarding left you without a project folder.
|
|
📎
|
Ops Agent workspace unavailable: what it means and fixes
Project → session → agent-home ladder. Company settings are the wrong surface. `/home/[]/...` is usually log redaction. Checklist to attach Local folder where the server runs. Read the post →
|
|
📎
|
Ops When board approvals leave agents hanging
The other “stuck but not thinking” pattern — decide + wake, not comment-only. Pairs with the workspace guide for silent stalls. Read the post →
|
Fresh listing
|
| Tool |
Paperclip AI Docker
Unofficial community Compose stack — build from source, agent CLIs preinstalled, bind 127.0.0.1, reverse-proxy ready. MIT. Not affiliated with upstream support.
|
|
|
📎
|
Pillar AI Agent Orchestration: How Agent Teams Actually Run
Heartbeats, atomic checkout, budgets before the call — the control-plane frame behind cwd and approvals. Read the post →
|
This week’s move: open the project on one coding agent (not company branding), set Local folder to an absolute path the Paperclip server can see, re-run a small issue, and confirm the fallback warning is gone. Self-hosting? Peek at the Docker listing before you invent compose from scratch — and treat product bugs as upstream, not the packager.
Ship something useful? Reply — next issue picks it up.
Cheers,
Aron 📎