메뉴
**When to use:** When building UI that displays session metrics (tokens, costs, agent stats) and needs to work for both active and inactive/historical sessions.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
How the multi-backend adapter system works — AgentAdapter interface, ACP protocol, session resume, and how to add new provider adapters
When debugging why an agent has the wrong provider/model, or when modifying agent creation or delegation flows.
How to get agents to use group chats for peer coordination in flightdeck-based crews. Covers hub-and-spoke anti-pattern, auto-grouping triggers, and file-lock-linked groups. Use when setting up multi-agent crews with 3+ agents or diagnosing why agents aren't collaborating directly.
When fetching from the `/coordination/activity` API and you only need specific action types (e.g., `progress_update`, `task_completed`, `delegated`).
When fetching activity data for UI display, or debugging why activity-based features show empty/stale data.
Architecture decisions and patterns from Flightdeck development (Phases 2–4). Covers feature architecture, state management, component patterns, and API design.
| name | live-data-db-fallback |
| description | **When to use:** When building UI that displays session metrics (tokens, costs, agent stats) and needs to work for both active and inactive/historical sessions. |
When to use: When building UI that displays session metrics (tokens, costs, agent stats) and needs to work for both active and inactive/historical sessions.
The AnalysisPage computed token totals from live in-memory agent objects:
const totalInput = currentAgents.reduce((s, a) => s + (a.inputTokens ?? 0), 0);
const totalOutput = currentAgents.reduce((s, a) => s + (a.outputTokens ?? 0), 0);
When a session is inactive, agents are gone from memory → all zeros → "Waiting for token data..."
// 1. Try live agent data first (real-time, most current)
let totalInput = currentAgents.reduce((s, a) => s + (a.inputTokens ?? 0), 0);
let totalOutput = currentAgents.reduce((s, a) => s + (a.outputTokens ?? 0), 0);
// 2. Fall back to persisted DB totals for inactive sessions
if (totalInput + totalOutput === 0) {
try {
const costsByAgent = await apiFetch<Array<{ totalInputTokens: number; totalOutputTokens: number }>>(
`/costs/by-agent?projectId=${projectId}`,
);
if (Array.isArray(costsByAgent)) {
totalInput = costsByAgent.reduce((s, c) => s + (c.totalInputTokens ?? 0), 0);
totalOutput = costsByAgent.reduce((s, c) => s + (c.totalOutputTokens ?? 0), 0);
}
} catch { /* costs API not available */ }
}
packages/web/src/components/AnalysisPage/AnalysisPage.tsx — token computation at ~line 84GET /costs/by-agent?projectId=X — defined in packages/server/src/routes/lead.tsCostTracker.getAgentCosts() in packages/server/src/agents/CostTracker.ts — queries task_cost_records table/costs/by-agent endpoint aggregates across ALL sessions for a project, not just the current one. This is fine for totals but could overcount if viewing a specific session.task_cost_records table is only populated when token usage events are recorded by AgentManager. If the dagTaskId gate blocks recording, the DB will also be empty (see the token-attribution-system skill).Array.isArray() on the response — the endpoint returns [] when CostTracker is unavailable.