fix(tui): stop ANSI escape leaks + drop redundant update_plan cards

[Vasanthdev2004]

What

Two chat-clarity fixes from comparing Zero's output to a reference agent on a "build a website" task.

1. ANSI escape garbage in the final answer 🐛

The final text was leaking raw, truncated terminal color codes as literal characters next to code blocks:

[1;4;38;2;236;
[38;2;236;236;238m    [m[38;2;

Root cause: styleAssistantMarkdownLine treated already-styled input (syntax-highlighted code, headings, tables — which carry real ANSI) byte-by-byte as runes and re-wrapped the whole line in another base.Render. That doubled the SGR density, and a downstream width-truncation then sliced mid-escape, dropping the leading ESC and printing the leftover ([38;2;…) as text. Only styled segments were affected — plain prose has no embedded escapes to split.

Fix: real escape sequences now pass through verbatim (using the existing ansiSequenceEnd helper, exactly like truncateStyledLine); the synthetic bold markers are still matched first so their style switch is preserved. ~14 lines.

2. Redundant update_plan tool cards 🧹

The plan is already shown in two places — the pinned plan panel above the composer and the PLAN sidebar — so rendering every update_plan call as its own transcript card was pure duplication (it stacked 3× in the demo). Generalized the existing Task-card skip into toolCardSuppressedInTranscript and now skip update_plan's call+result rows too. The plan-panel sync and session events are unchanged.

Tests

• TestStyleAssistantMarkdownLinePassesAnsiVerbatim — proves already-styled input passes through verbatim; verified it fails before the fix (output re-wraps the input escape inside base's \x1b[1m).
• TestToolCardSuppressedInTranscript — Task/update_plan suppressed, everything else still shown.
• Full ./internal/tui/... suite green; build + gofmt clean.

Part of a series

This is 1 of 3 from the chat-clarity comparison. Follow-ups: diff/code preview on write/edit cards (the render half + go-udiff already exist), and a system-prompt preamble so the agent briefly states its approach. Independent PRs.

Summary by CodeRabbit

• Bug Fixes
  • Improved Markdown styling in the TUI so pre-styled text with ANSI colors renders correctly without broken or duplicated formatting.
  • Reduced duplicate transcript entries by hiding tool cards that are already shown in other parts of the interface, including plan-related and specialist activity.

[github-actions]

Zero automated PR review

Verdict: No blockers found

Blockers

• None found.

Validation

• [pass] Diff hygiene: git diff --check
• [pass] Tests: go test ./...
• [pass] Build: go run ./cmd/zero-release build
• [pass] Smoke build: go run ./cmd/zero-release smoke

Scope

Head: aa3aa4ba8db4
Changed files (3): internal/tui/assistant_markdown.go, internal/tui/model.go, internal/tui/rendering_lime_test.go

This deterministic review checks validation status and basic diff hygiene. A human reviewer still owns product judgment and design quality.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 635b26c1-5244-4179-8f4a-17b74740c30f

📥 Commits

Reviewing files that changed from the base of the PR and between 46c44a8 and aa3aa4b.

📒 Files selected for processing (3)

• internal/tui/assistant_markdown.go
• internal/tui/model.go
• internal/tui/rendering_lime_test.go

---

Walkthrough

This PR preserves ANSI escape sequences in assistant markdown rendering and centralizes transcript-card suppression for Task and update_plan tool events, with regression tests covering both behaviors.

Changes

ANSI assistant markdown styling

Layer / File(s)	Summary
ANSI escape passthrough internal/tui/assistant_markdown.go, internal/tui/rendering_lime_test.go	styleAssistantMarkdownLine now writes detected ANSI escape sequences verbatim after flushing buffered text, and the regression test checks ANSI token and reset preservation.

Tool transcript suppression

Layer / File(s)	Summary
Transcript suppression helper internal/tui/model.go, internal/tui/rendering_lime_test.go	toolCardSuppressedInTranscript returns true for Task and update_plan, and both tool-call and tool-result transcript rows use it before emitting cards.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~15 minutes

Possibly related PRs

• Gitlawb/zero#187: Modifies internal/tui/assistant_markdown.go, the same styling path extended here to pass ANSI escape sequences through verbatim.
• Gitlawb/zero#83: Touches internal/tui/model.go tool transcript handling, which this PR updates to suppress Task and update_plan cards.

Suggested reviewers

• gnanam1990

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes both main changes: ANSI escape passthrough fixes and redundant update_plan transcript card removal.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

---

_{Comment @coderabbitai help to get the list of available commands.}