UNPKG

claude-flow

Version:

Ruflo - Enterprise AI agent orchestration for Claude Code. Deploy 60+ specialized agents in coordinated swarms with self-learning, fault-tolerant consensus, vector memory, and MCP integration

33 lines 1.32 kB
/** * GAIA CodeAgent Smoke Tests — ADR-138 iter 54 (FINAL design) * * 5 tests validating the CodeAgent harness end-to-end. * Tests 1-4 are unit-level (no API calls). Test 5 uses 1 API call. * * Tests: * T1: Code block extraction — parser correctness (no subprocess) * T2: Python runner — simple math (subprocess, no API) * T3: Python runner — file read via ATTACHMENT_PATH * T4: Python runner — code error recovery (traceback returned as obs) * T5: End-to-end — simple factual question (1 API call, ~$0.01) * * Pass criteria: 5/5. Gate before iter 55 5Q run. * * Run: npx tsx gaia-codeagent.smoke.ts * or: node dist/src/benchmarks/gaia-codeagent.smoke.js * * Refs: ADR-138, #2156, iter 54 */ interface TestResult { name: string; passed: boolean; message: string; durationMs: number; } declare function testCodeBlockExtraction(): TestResult; declare function testPythonRunnerMath(): TestResult; declare function testPythonRunnerFileRead(): TestResult; declare function testPythonRunnerErrorRecovery(): TestResult; declare function testEndToEndSimple(): Promise<TestResult>; export { testCodeBlockExtraction, testPythonRunnerMath, testPythonRunnerFileRead, testPythonRunnerErrorRecovery, testEndToEndSimple, }; //# sourceMappingURL=gaia-codeagent.smoke.d.ts.map