feat: fuzzy substring search with ranked scoring for slash commands

[felipeggv]

Summary

Replaces the prefix-only slash command filter with ranked substring matching. Users can now find commands by typing any part of the command name or description — similar to VS Code's Command Palette.

Problem

The slash command dropdown only matched commands that started with the typed text. With growing numbers of custom commands (Spec-Kit, OpenSpec, user-defined), users had to remember exact command prefixes to find anything.

Example: typing /prd returned nothing, even though /GLOBAL:agents:prd-creator exists.

Solution

Three-tier ranked scoring system applied to both InputArea.tsx (dropdown render) and App.tsx (keyboard handler):

Score	Match Type	Example (input: hist)
3	Prefix on command name	/history → appears first
2	Substring in command name	/command-history-viewer → appears second
1	Substring in description	/clear (if description mentions "history") → appears last

Results are sorted descending by score. The leading / is stripped from both input and command name before comparison to avoid false negatives.

Before / After

User types	Before	After
/prd	No results	/GLOBAL:agents:prd-creator (score 2)
/synopsis	No results	/history (description contains "synopsis", score 1)
/ard	No results	/wizard (name contains "ard", score 2)
/hist	/history only	/history (score 3, prefix) + description matches (score 1)
/clear	/clear only	/clear (score 3, prefix) — same behavior, backwards compatible

How This Helps Users

• Discoverability — Find commands by what they do, not just their exact name
• Partial recall — Don't need to remember full command prefixes
• Scales with custom commands — As users add Spec-Kit, OpenSpec, and custom slash commands, search stays useful
• No noise — Ranked scoring ensures prefix matches always appear above description-only matches

Files Changed

File	Change
src/renderer/components/InputArea.tsx	Dropdown filter: startsWith → scored includes on name + description
src/renderer/App.tsx	Keyboard handler filter: same scoring logic

Implementation Details

// Scoring logic (same in both files)
const cmdName = cmd.command.toLowerCase().replace(/^\//, '');
if (cmdName.startsWith(searchTerm))  → score 3 (prefix match)
if (cmdName.includes(searchTerm))    → score 2 (substring in name)
if (description.includes(searchTerm)) → score 1 (substring in description)
// No match → excluded from results

• searchTerm strips the leading / from user input
• cmd.description is guarded with && (some custom commands may not have descriptions)
• useMemo dependency updated from inputValueLower to searchTerm in InputArea

Test Plan

• Type / — all commands appear (empty search term = score 0, all included)
• Type /prd — prd-creator command appears
• Type /synopsis — /history appears (description match)
• Type /hist — /history appears first (prefix match before description matches)
• Type /s — prefix matches (e.g., /status) appear above description-only matches
• Arrow keys navigate filtered list correctly
• Enter/Tab selects the highlighted command
• Escape closes the dropdown
• Terminal-only commands don't appear in AI mode (mode filtering unchanged)
• npm run lint — zero new errors
• npm test -- --run — all tests pass (5 pre-existing failures in symphony.test.ts)
• npm run build — clean production build

🤖 Generated with Claude Code

[pedramamini]

@CodeRabbit review

[coderabbitai]

✅ Actions performed

Review triggered.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[coderabbitai]

Warning

Rate limit exceeded

@pedramamini has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 12 minutes and 54 seconds before requesting another review.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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