feat(cli): add /brief for terse-response mode

[emal-avala]

Summary

A new /brief slash command that toggles brief mode — a session-local flag that instructs the model to keep responses under 3 sentences unless the task genuinely needs more.

When on, an instruction block is injected near the top of the system prompt (before the tool docs) so it stays salient in long contexts:

# Response style

Brief mode is enabled. Keep responses under 3 sentences unless the user
asks for detail or the task genuinely requires a longer answer ...

Why

During long pair-programming sessions the model's default "paragraphs with headers and a recap" style creates friction. The user wants a one-line way to say "cool it" without editing the system prompt config. Not persisted — brief mode resets on a new session.

Behaviour

• /brief when off → Brief mode enabled. Responses will be kept terse (≤3 sentences).
• /brief when on → Brief mode disabled. Response style restored.
• System prompt gains/loses a # Response style block accordingly
• Block appears before # Available Tools so it's not buried

Test plan

• cargo fmt --all — clean
• cargo clippy --workspace --all-targets -- -D warnings — clean
• cargo test -p agent-code-lib --lib state — 5/5 pass (existing state tests updated for new field)
• cargo test -p agent-code-lib --lib system_prompt — 4/4 pass (2 new)
  • system_prompt_omits_brief_block_when_off
  • system_prompt_injects_brief_block_when_on (also verifies position)

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.