Consider it done. The open-source AI agent that works out of the box · 想到,就能做到。开源、开箱即用的 AI Agent。
TypeScript agentai-agentai-assistantandroidclaude-codecodexelectroniosllmmacosreact-nativetypescriptwindows
import { describe, expect, it } from 'vitest';
import {
buildContextUsageCreateOpts,
canUseLocalCodexRateLimitControl,
shouldFallbackToLegacyCodexUsage,
summarizeContextUsage,
summarizeSessionSpend,
} from '@/session/sessionControls';
import type { RemoteSession } from '@/session/types';
function session(patch: Partial<RemoteSession> = {}): RemoteSession {
return {
id: 's1',
userId: 'u1',
title: 'Session',
workingDir: '/repo',
workspaceKind: 'project',
model: 'claude-sonnet-4-6',
effort: 'medium',
permissionMode: 'ask',
fastMode: false,
status: 'active',
agentKind: 'cc',
userSendAt: null,
createdAt: '2026-01-01T00:00:00.000Z',
updatedAt: '2026-01-01T00:00:00.000Z',
...patch,
};
}
describe('sessionControls', () => {
it('builds desktop-compatible context usage create options', () => {
expect(buildContextUsageCreateOpts(session())).toEqual({
agentKind: 'claude-code',
workingDir: '/repo',
model: 'claude-sonnet-4-6',
effort: 'medium',
permissionMode: 'ask',
fastMode: false,
});
expect(buildContextUsageCreateOpts(session({ agentKind: 'codex', fastMode: true }))).toMatchObject({
agentKind: 'codex',
fastMode: true,
});
});
it('exposes local Codex quota controls only for local subscription sessions', () => {
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'gpt-5.5',
}))).toBe(true);
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'gpt-5.5',
providerId: 'openai',
}))).toBe(true);
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'gpt-5.5',
remoteHostId: 'ssh-host-1',
}))).toBe(false);
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'gpt-5.5',
providerId: 'xd',
}))).toBe(false);
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'codex/gpt-5.5',
}))).toBe(false);
expect(canUseLocalCodexRateLimitControl(session({
agentKind: 'codex',
model: 'xai/grok-4.3',
providerId: 'xai',
}))).toBe(false);
expect(canUseLocalCodexRateLimitControl(session({ agentKind: 'cc' }))).toBe(false);
});
it('does not reuse legacy quota after an account-change precondition', () => {
expect(shouldFallbackToLegacyCodexUsage({
code: 'PRECONDITION_FAILED',
message: 'ACCOUNT_CHANGED: refresh usage',
})).toBe(false);
expect(shouldFallbackToLegacyCodexUsage(new Error('new channel unavailable'))).toBe(true);
});
it('summarizes common context usage shapes for mobile display', () => {
expect(summarizeContextUsage({
contextTokens: 12000,
maxContextTokens: 200000,
percent: 0.06,
costUsd: 0.12345,
})).toEqual({
title: 'Context usage',
detail: '12,000 / 200,000 tokens · 6% · $0.1235',
rows: [],
});
// Regression: Claude SDK reports `percentage` on a 0–100 scale, unlike legacy `percent`.
expect(summarizeContextUsage({ percentage: 1 })).toMatchObject({
detail: '1%',
});
expect(summarizeContextUsage(null)).toEqual({
title: 'Context usage',
detail: '暂无上下文数据',
rows: [],
});
});
it('summarizes desktop structured context usage details for mobile system cards', () => {
expect(summarizeContextUsage({
totalTokens: 42000,
rawMaxTokens: 200000,
percentage: 21,
model: 'claude-sonnet-4-6',
categories: [
{ name: 'System prompt', tokens: 10000, color: '#111' },
{ name: 'MCP tools', tokens: 8000, color: '#222' },
{ name: 'Free space', tokens: 158000, color: '#ddd' },
],
mcpTools: [
{ serverName: 'github', name: 'search_issues', tokens: 3000 },
{ serverName: 'browser', name: 'open', tokens: 1000, isLoaded: false },
],
memoryFiles: [{ path: '/repo/MEMORY.md', type: 'project', tokens: 1200 }],
agents: [{ agentType: 'reviewer', source: 'project', tokens: 900 }],
skills: {
includedSkills: 2,
totalSkills: 5,
tokens: 2400,
skillFrontmatter: [{ name: 'review-pr', source: 'user', tokens: 600 }],
},
slashCommands: { includedCommands: 3, totalCommands: 9, tokens: 700 },
messageBreakdown: {
userMessageTokens: 500,
assistantMessageTokens: 600,
toolCallTokens: 700,
toolResultTokens: 800,
attachmentTokens: 0,
toolCallsByType: [{ name: 'Read', callTokens: 100, resultTokens: 200 }],
attachmentsByType: [],
},
apiUsage: {
input_tokens: 1000,
cache_creation_input_tokens: 200,
cache_read_input_tokens: 300,
output_tokens: 400,
},
})).toMatchObject({
title: 'Context usage',
detail: '42,000 / 200,000 tokens · 21%',
rows: expect.arrayContaining([
{ label: 'model', value: 'claude-sonnet-4-6' },
{ label: 'System prompt', value: '10,000 tokens · 5.0%' },
{ label: 'MCP tools', value: expect.stringContaining('github/search_issues') },
{ label: 'Memory files', value: expect.stringContaining('MEMORY.md · project') },
{ label: 'Skills', value: expect.stringContaining('2 / 5') },
{ label: 'Slash commands', value: expect.stringContaining('3 / 9') },
{ label: 'Tool types', value: expect.stringContaining('Read') },
{ label: 'API output', value: '400 tokens' },
]),
});
});
it('summarizes session spend from desktop session usage fields', () => {
expect(summarizeSessionSpend(session({
totalCostUsd: 0.0342,
totalTokenUsage: 12450,
contextTokens: 6000,
contextWindow: 200000,
}))).toEqual({
title: 'Session spend',
detail: '本任务 $0.03 · 12.5k tokens · 上下文 6k / 200k · 3%',
available: true,
});
expect(summarizeSessionSpend(session())).toEqual({
title: 'Session spend',
detail: '暂无任务用量',
available: false,
});
});
});