refactor: extract getStdinFlags helper for Windows stdin transport

[pedramamini]

Summary

• Extracts duplicated Windows stdin flag logic into a shared getStdinFlags() utility
• Adds SSH session exclusion to useInputProcessing and inlineWizardConversation paths (previously missing)
• Follows up on Fix local Auto Run Codex prompt transport on Windows #376

Changes

• New: src/renderer/utils/spawnHelpers.ts — single source of truth for sendPromptViaStdin / sendPromptViaStdinRaw computation
• Updated: useAgentExecution.ts, useInputProcessing.ts, inlineWizardConversation.ts — replaced 4 inline copies with getStdinFlags() calls
• New: src/__tests__/renderer/utils/spawnHelpers.test.ts — 5 tests covering all flag combinations

Context

PR #376 added stdin transport flags to the batch spawn paths in useAgentExecution. This brought the total to 4 copies of the same 6-line computation across the renderer. This PR consolidates them and makes SSH exclusion consistent everywhere (the spawner's if-else chain already made this safe, but the explicit guard is clearer).

Test plan

• npm run lint passes
• All 71 affected tests pass (spawnHelpers, useAgentExecution, useInputProcessing, inlineWizardConversation)
• Windows Auto Run behavior unchanged (stdin transport still active for local runs)
• SSH sessions unaffected (flags correctly disabled)

[claude]

Code Review - PR #395

Summary

This is an excellent refactoring that consolidates duplicated Windows stdin transport logic into a single, well-tested utility function. The PR successfully addresses code duplication while improving consistency and maintainability.

---

✅ Strengths

1. Clean Abstraction

   • The getStdinFlags() utility in src/renderer/utils/spawnHelpers.ts:11-22 provides a single source of truth for stdin transport logic
   • Clear JSDoc explaining the Windows cmd.exe limitation and SSH exclusion rationale
   • Simple, focused interface with explicit input parameters

2. Comprehensive Test Coverage

   • All 5 test cases cover the complete decision matrix:
     • Non-Windows platforms
     • Windows with/without stream-json support
     • SSH session exclusions (2 test cases)
   • Tests use proper mocking of navigator.platform

3. Bug Fix: SSH Session Handling

   • Previously, useInputProcessing.ts and inlineWizardConversation.ts were missing SSH session checks
   • Now all 4 call sites consistently exclude SSH sessions from stdin transport
   • This prevents potential conflicts with SSH's dedicated stdin-script handling

4. Consistent Implementation

   • All refactored sites (useAgentExecution.ts, useInputProcessing.ts, inlineWizardConversation.ts) now follow the same pattern
   • Reduced from ~6 lines per site to a single function call with clear parameters

---

🔍 Code Quality Observations

Good:

• ✅ Follows DRY principle effectively
• ✅ Maintains backward compatibility (behavior unchanged for existing flows)
• ✅ Clear variable naming (sendPromptViaStdin vs sendPromptViaStdinRaw)
• ✅ Proper SSH exclusion handling per [[CLAUDE-PATTERNS.md]]

Minor Notes:

• The navigator.platform check is renderer-specific (correct for this context, but worth noting this utility is renderer-only)
• Consider future enhancement: platform detection could potentially be abstracted further if main process ever needs similar logic (not needed now)

---

🔒 Security & Performance

Security: ✅ No concerns

• Stdin transport is a safer alternative to command-line arguments (avoids shell injection via special characters)
• SSH exclusion properly maintained

Performance: ✅ No concerns

• Negligible overhead (simple boolean checks)
• No new IPC calls or async operations

---

🧪 Testing

Test Coverage: ✅ Excellent

• New test file with 5 comprehensive test cases
• All edge cases covered (platform × SSH × stream-json support)
• Proper cleanup with afterEach restore

Suggested Test Verification:

npm run test -- src/__tests__/renderer/utils/spawnHelpers.test.ts
npm run lint

---

📋 Recommendations

Must Do (Before Merge):

• Verify all existing tests pass (71 tests mentioned in test plan)
• Confirm npm run lint passes
• Manual smoke test: Windows Auto Run behavior unchanged
• Manual smoke test: SSH sessions still work correctly

Nice to Have (Future):

• Consider adding a comment in CLAUDE-PLATFORM.md referencing this utility for future Windows platform work
• If more renderer utilities emerge, consider creating a src/renderer/utils/platform.ts module

---

✅ Verdict

LGTM with approval pending test verification.

This PR demonstrates excellent software engineering:

• Identifies and eliminates code duplication
• Adds tests for the new abstraction
• Fixes previously missed SSH exclusions
• Maintains clean, readable code

The refactoring is low-risk (pure consolidation) and high-value (improves maintainability and fixes edge cases).

Risk Assessment: 🟢 Low

• No behavior changes for existing paths
• New utility is well-tested
• Follows established patterns per CLAUDE.md

Once tests pass, this is ready to merge.

---

Review conducted per CLAUDE.md guidelines, focusing on code quality, security, performance, and test coverage.