feat: Add support for opencode.ai

[dragon-ai-agent]

This commit adds support for the opencode.ai CLI as a new coder in metacoder.

A new OpencodeCoder class has been implemented to wrap the opencode run command.

The coder has been registered in the application and a test case has been added to ensure it works as expected.

[cmungall]

@claude review

[claude]

Claude finished @dragon-ai-agent's task —— View job

---

PR Review: Add support for opencode.ai

Review Checklist:

• Read and analyze changed files
• Review OpencodeCoder implementation
• Check registry integration
• Review test coverage
• Assess code quality and best practices
• Check for potential bugs or issues
• Evaluate performance considerations
• Review security concerns
• Provide final feedback

---

Overall Assessment

This PR successfully adds support for opencode.ai CLI following the established patterns in the codebase. The implementation is clean and consistent with existing coders.

Code Quality ✅

Positives:

• Follows existing patterns perfectly (matches claude.py structure almost exactly)
• Clean separation of concerns with OpencodeConfig class
• Proper inheritance from BaseCoder
• Good error handling with try/catch blocks
• Consistent logging and command construction
• Proper type annotations throughout

Implementation Analysis

src/metacoder/coders/opencode.py - Good

• Pattern consistency: Excellent adherence to existing coder patterns
• Configuration: Proper Pydantic model with sensible defaults
• Command construction: Follows established pattern with proper shell escaping
• Error handling: Appropriate FileNotFoundError and generic exception handling
• Logging: Consistent with other implementations

src/metacoder/registry.py - Good

• Clean integration following existing pattern
• Proper import and registration

tests/test_coders/test_coders_basic.py - Good

• Added to existing test structure appropriately
• Tests basic initialization and string representation

Minor Issues & Suggestions

1. Documentation (Minor)

# Current in opencode.py:21
model: str = Field(default="openai/o1-mini", description="Model to use for OpenCode")

# Consider more descriptive:
model: str = Field(default="openai/o1-mini", description="Model identifier for OpenCode (e.g., 'openai/o1-mini', 'anthropic/claude-3-sonnet')")

2. Command Flag Consistency (Minor)

The implementation uses --task for opencode vs --prompt for claude. This appears correct based on the respective CLI interfaces, but worth documenting if opencode.ai CLI indeed uses --task.

3. Process Handling (Very Minor)

The subprocess handling is identical to claude.py, which is good for consistency. Both could potentially benefit from timeout handling, but that's beyond scope of this PR.

Security Analysis ✅

• Proper use of shlex.quote() for command construction
• No shell injection vulnerabilities
• No hardcoded credentials or sensitive data
• Safe subprocess execution

Performance Considerations ✅

• Efficient streaming output handling
• No blocking operations in async context
• Minimal memory footprint
• Subprocess cleanup handled properly

Test Coverage 📝

Current coverage is adequate for basic functionality:

• ✅ Initialization testing
• ✅ String representation testing
• ❓ Missing: Interactive mode testing (but this is consistent with other coders in the test file)

Note: The test approach matches existing coders - no interactive mode testing in basic tests, likely covered elsewhere or requiring integration tests.

Potential Issues

1. Dependency Documentation (Minor)

Consider documenting the opencode.ai CLI installation requirement in the project docs.

2. Model Validation (Future Enhancement)

No validation of model string format, but this is consistent with other coders.

Recommendations

1. Documentation: Add brief comment about opencode.ai CLI requirement
2. Future: Consider adding integration tests for actual CLI interaction (applies to all coders)

Final Verdict: ✅ APPROVED

This is a high-quality implementation that:

• Follows established patterns perfectly
• Maintains code consistency across the project
• Includes appropriate tests
• Has no security or performance issues
• Integrates cleanly with existing architecture

The code is ready to merge. Great work maintaining consistency with the existing codebase!

---