feat(agents): rich multi-agent status — GUI, island, Mac menu bar (Phase 1)#107
Merged
Conversation
…r (Phase 1)
Layer 1 of the agent control plane (plan: iridescent-marinating-torvalds). The
observers already capture rich progress (turns/tokens/tools/files/branch/
pendingPermission); this surfaces it everywhere instead of pip+name+time.
- agent-roster.ts: new pure, source-injected `formatActivity(session)` — one-line
"⚠pending | ✗err · turn N · Mk tok · $cmd · Tool file". Replaces the island's
local copy (+ its basename helper) so island + sidebar share one tested impl
(injection-safety test + unit tests cover it).
- island: inject ${formatActivity}; header "claude code"→"agents"; cap 5→8.
- GUI sidebar (lisa-client/lisa-css/lisa-html): a second per-row activity line
(inline sbActivity mirrors the shared fn — this client script is a
no-interpolation template literal), "agents" header, cap 5→8.
- Mac menu bar (MenuBarController): poll /api/agents/sessions (was claude-only),
aggregate state counts across ALL agents, and the popover now lists per-agent
rows (state dot + kind + branch label + activity line). New DTOs + Swift
agentLabel/agentActivity mirroring the web helpers.
700 tests pass (+5); typecheck + build clean; swift build clean; dist island
verified (formatActivity injected, no __name, basename removed).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Layer 1 of the approved agent control plane (
plans/iridescent-marinating-torvalds.md): make every claude-code / codex / … agent's progress + status visible. The observers already capture richSessionActivity(turns, tokens, last tool·file, branch, pendingPermission); the UI just showed pip + name + time. This surfaces it everywhere.How
agent-roster.ts— new pure, source-injectedformatActivity(session):⚠ pending | ✗ err · turn N · Mk tok · $ cmd · Tool file. Replaces the island's local copy (+ itsbasename), so island + sidebar share one tested impl (injection-safety + unit tests).${formatActivity}; header "claude code" → "agents"; row cap 5 → 8.lisa-client/lisa-css/lisa-html) — a second per-row activity line (inlinesbActivitymirrors the shared fn, since this client script is a documented no-interpolation template literal); "agents" header; cap 5 → 8.MenuBarController) — poll/api/agents/sessions(was claude-only); aggregate state counts across all agents; the popover now lists per-agent rows (state dot + kind + branch label + activity line). New DTOs + SwiftagentLabel/agentActivitymirroring the web helpers.Verification
formatActivityrules + injection-safety;ISLAND_HTML/MAIN_HTMLparse; snapshot updated)npm run typecheck+npm run buildclean;swift buildclean; dist island verified (formatActivityinjected, no__name,basenameremoved)Next (per plan)
Phase 2 (Stage A): per-agent Cancel button +
/api/agents/dispatch+ prominent pendingPermission. Phase 3 (Stage B): managed agents — the real "send commands to a running agent."🤖 Generated with Claude Code