test: add test coverage for `rafter brief` command

[Rome-1]

What

The `rafter brief` command has no automated tests in either implementation. Add comprehensive test coverage.

Context

`rafter brief` delivers platform-specific security knowledge to AI agents. It supports 8 platforms and multiple topics. See the existing `report.test.ts` (219 lines, 16 test cases) as a reference for test style.

What to test

• All topic listing and help output
• Each platform-specific guide renders correctly
• Agent mode output (plain text, no colors/emoji)
• Invalid topic handling (graceful error)
• Help flag output

Files to create

1. `node/tests/brief.test.ts` — Vitest tests (reference: `node/tests/report.test.ts`)
2. `python/tests/test_brief.py` — pytest tests

Commands

• `node/src/commands/brief.ts` — Node implementation
• `python/rafter_cli/commands/brief.py` — Python implementation

How to run

```bash
cd node && npx vitest run tests/brief.test.ts
cd python && pytest tests/test_brief.py -v
```

[pedropachecog]

Hi! I would like to work on this issue as part of CodePath AI301. I noticed that node/tests/brief.test.ts and python/tests/test_brief.py already exist on main, so my first step will be to verify which branches from the original checklist still need coverage, especially agent-mode/plain-text output and any parity gaps between the Node and Python tests. If there is a specific remaining gap you want prioritized, I would appreciate the pointer.

[Rome-1]

Thanks for picking this up, and welcome! You're reading it right — node/tests/brief.test.ts and python/tests/test_brief.py both already landed on main and cover most of the checklist (topic listing, per-platform setup guides, the CYOA skill hierarchy, and error handling). So rather than redo that, here are the two genuine gaps worth prioritizing:

1. Agent-mode / plain-text output — the top gap. There's a real global -a/--agent flag ("Plain output for AI agents (no colors/emoji)", node/src/index.ts), but neither suite asserts that rafter brief <topic> --agent output is actually plain — no ANSI escape sequences, no emoji leaking through. A couple of assertions locking that contract in (plus the Python equivalent) closes the checklist's "Agent mode output" item. Exactly the branch you flagged — good instinct.

2. Node↔Python parity on the internal helpers. The Python suite has a TestInternalFunctions class (frontmatter stripping, extract_sections edge cases — code blocks, case-insensitivity, no-match — plus render_topic/unknown-platform returns and topic-list formatting) that the Node suite has no equivalent for. Right now brief.ts only exports createBriefCommand, so those helpers aren't unit-testable from Node yet — closing this likely means exporting a few internal functions and adding matching Vitest cases. Well-scoped and high-value.

If you want to keep it small, (1) is a clean self-contained PR and (2) is a natural follow-up — either is welcome. One ask: both implementations must stay in lockstep, so please mirror whatever you add in both Node and Python. Looking forward to the PR!