test: add CLI test coverage (39 tests)

[robotlearning123]

Summary

• Refactored cli.ts to export program and guard parseAsync() with isMain check (+3 lines) for testability
• Added src/__tests__/cli.test.ts with 39 BDD-style tests covering all 11 CLI commands
• Total test suite: 216 → 255 tests, all passing

Coverage

• All 11 commands: submit, batch, list, status, cancel, retry, diff, logs, stats, workers, search
• Global flags: --json (parameterized across 6 commands), --url
• Error handling: HTTP 404/500, network errors (ECONNREFUSED)
• API behavior: Content-Type headers, request body construction, URL query params, output formatting

Approach

Mock globalThis.fetch to verify HTTP method + URL + body for each command. Capture console.log to verify output formatting (table vs JSON). No subprocess spawning needed.

Test plan

• npx tsc --noEmit — zero errors
• node --import tsx --test src/__tests__/cli.test.ts — 39/39 pass
• Full suite src/__tests__/*.test.ts — 255/255 pass
• Pre-commit hooks pass on both commits

🤖 Generated with Claude Code

[robotlearning123]

Code review

Found 1 issue:

1. CLAUDE.md not updated to reflect the new test suite. The "Build, Test & Run" section says "217 tests across 5 suites" and the "Test Files" section lists exactly 5 files. This PR adds a 6th suite (cli.test.ts) with 39 tests (255 total across 6 suites) but does not update either section. Per CLAUDE.md: "CLAUDE.md is the agent's single source of truth, enforced mechanically."

cc-manager/CLAUDE.md

Lines 30 to 35 in 7e9222c

	```bash
	# Run tests (217 tests across 5 suites)
	node --import tsx --test src/__tests__/*.test.ts
	```

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}

[robotlearning123]

Code review

No issues found. Previous issue (CLAUDE.md test count/file listing) has been fixed in ab8e9e5. Re-checked for bugs and CLAUDE.md compliance.

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}