You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I ran squad init in a new Squad scaffold and asked the coordinator to hire a team. The setup took about 11 minutes (user-reported) before a generic team was usable.
The friction was cumulative:
squad init exited successfully but left .squad/team.md with an empty ## Members table. The coordinator setup skill says to enter Init Mode when team.md is absent, so this initialized-but-empty state can miss the fresh-team flow. Detect an empty Members table as uninitialized too.
Falling back to the checked-in casting templates did not work with the installed validator: the registry template lacks required display_name, role, and updated_at fields, and the history snapshot shape also differs. squad health --json then failed with registry contains invalid agent records; manual registry/history and commit-manifest repair was needed.
After manually adding the roster, specialist charters, and routes, squad health --json and squad doctor passed.
Suggested follow-up
Make the post-init path detect an empty roster and lead directly to a working team-creation flow. Ensure the shipped casting templates match the current registry/history validator, and add a fresh-directory regression test that runs init, creates the first specialist, and finishes with squad health/squad doctor passing. Until cast --name/--role is implemented, its help and exit status should not imply successful creation.
This report is about the end-to-end first-run experience; #2191 already tracks the cast command no-op itself.
Thank you for the report and for sharing the workaround, @Omzig — removing tools: ["*"] from the squad.agent.md front matter was exactly the right diagnosis. I'm closing this as stale since it was filed against an older build, and the later first-run friction you described deserves its own focused issue. If you can still reproduce this on a current version, please open a new issue with fresh repro steps and we'll jump on it. Really appreciate the detail! 💛
Activity
Omzig commentedon Oct 16, 2026
remove
tools: ["*"]in the front matter of squad.agent.md[-]squad.agent.md tools * in the new vscode causes problems[/-][+]cannot hire a team using squad[/+][-]cannot hire a team using squad[/-][+]cannot hire a team after squad init[/+]Omzig commentedon Oct 16, 2026
Fresh first-run reproduction (Windows, Squad CLI 1.0.0)
I ran
squad initin a new Squad scaffold and asked the coordinator to hire a team. The setup took about 11 minutes (user-reported) before a generic team was usable.The friction was cumulative:
squad initexited successfully but left.squad/team.mdwith an empty## Memberstable. The coordinator setup skill says to enter Init Mode whenteam.mdis absent, so this initialized-but-empty state can miss the fresh-team flow. Detect an empty Members table as uninitialized too.squad cast --name Lead --role leadprintedfull implementation pending, exited successfully, and created nothing. This is the same no-op behavior documented in Bug: cast --name/--role and hire are no-op stubs but exit successfully (v1.0.1) #2191.display_name,role, andupdated_atfields, and the history snapshot shape also differs.squad health --jsonthen failed withregistry contains invalid agent records; manual registry/history and commit-manifest repair was needed.squad health --jsonandsquad doctorpassed.Suggested follow-up
Make the post-init path detect an empty roster and lead directly to a working team-creation flow. Ensure the shipped casting templates match the current registry/history validator, and add a fresh-directory regression test that runs init, creates the first specialist, and finishes with
squad health/squad doctorpassing. Untilcast --name/--roleis implemented, its help and exit status should not imply successful creation.This report is about the end-to-end first-run experience; #2191 already tracks the cast command no-op itself.
fix(cli): 2.0.0-preview.2 upgrade/health fixes from dogfooding
fix(cli): 2.0.0-preview.2 upgrade/health fixes from dogfooding (#2303)
bradygaster commentedon Nov 5, 2026
Thank you for the report and for sharing the workaround, @Omzig — removing
tools: ["*"]from thesquad.agent.mdfront matter was exactly the right diagnosis. I'm closing this as stale since it was filed against an older build, and the later first-run friction you described deserves its own focused issue. If you can still reproduce this on a current version, please open a new issue with fresh repro steps and we'll jump on it. Really appreciate the detail! 💛