fix(cli): classify broken pipes as user aborts

[cursor]

Summary

• preserve Rich's closed-pipe handling while exposing a distinct CLI exception
• classify closed output pipes as user aborts instead of command failures
• preserve the existing exit status and add console/lifecycle regression coverage

The daily CLI failure export contained two current-version endpoints events failures that completed all API requests and then reported only status 1. Rich 14.2 converts BrokenPipeError to SystemExit(1), which made expected early pipe closure look like a CLI defect.

Verification

• TOGETHER_TELEMETRY_DISABLED=1 uv run pytest tests/cli/test_command_telemetry.py tests/cli/test_console_theme.py -n 0 (24 passed)
• ./scripts/lint

  

[broly-code-security-scanner]

Broly Security Scan

Warning

⚠️ Potential results — Broly is still completing analysis or verification. This comment will update automatically.
Files pending AI analysis: 4; findings pending verification: 1.

Warning

Summary

1 actionable finding(s) in this PR

• 🔴 1 critical

All actionable items are in the table below.

Severity	Scanner	Issue	Location	Dismiss
🔴 CRITICAL	SECRETS	Hardcoded secret	src/together/lib/cli/init.py:160	d1

Dismiss false positives

Important

Secrets cannot be dismissed with a checkbox. Broly decides whether a credential is real or a test fixture using AI, so overriding that verdict means adding evidence it did not have — reply with /broly dismiss KEY: your reason.
If the same fixture pattern keeps getting flagged, declare it once as redaction_skip_prefixes in .broly.yaml instead of dismissing it each time.

• 🔑 d1 · 🔴 CRITICAL · src/together/lib/cli/__init__.py:160 · Hardcoded secret — reply /broly dismiss d1: your reason

Note

Re-scan this PR anytime with /broly scan — useful after /broly undismiss, or to refresh findings without a new push.

Broly — SAST (zai-org/GLM-5.3-Flash) · Secrets · SCA · IaC · GH Actions · Base Images · Supply Chain Threats · Exploit Chains · Adversarial Verification

We're continuously improving Broly's accuracy and finding quality — your feedback is valuable. False positives, missed findings, bugs, and feature requests all welcome.