Guide for building professional, high-accuracy algorithms with a focus on documentation verification and error-free execution. Use this skill when the user asks for "algorithms", "complex logic", "robust functions", or explicitly mentions "no errors" and "check docs".
Installs into .claude/skills of the current project.
Are you the author of Algorithm Builder?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/david-li0406-algorithm-builder)
---
name: algorithm-builder
description: Guide for building professional, high-accuracy algorithms with a focus on documentation verification and error-free execution. Use this skill when the user asks for "algorithms", "complex logic", "robust functions", or explicitly mentions "no errors" and "check docs".
---
# Algorithm Builder
This skill enforces a disciplined approach to creating professional algorithms, ensuring accuracy through documentation verification and robust error handling.
## Workflow
### 1. Research & Verification (MANDATORY)
**Before writing any code:**
1. **Consult Documentation**: Use `context7` or `search_web` to find the official documentation for relevant libraries (e.g., standard libraries, third-party packages).
- _Do not guess_ API signatures or behaviors.
- _Do not rely_ solely on pre-training knowledge for critical logic.
2. **Verify Assumptions**: Confirm edge case behaviors (e.g., empty inputs, nulls, boundary values) from authoritative sources.
### 2. Design & Strategy
1. **Pseudocode**: Briefly outline the algorithm steps in comments or a scratchpad.
2. **Edge Case Identification**: List at least 3 potential failure modes or edge cases (e.g., "Input array is empty", "Network timeout", "Invalid date format").
3. **Complexity Analysis**: Briefly note time/space complexity if ensuring performance is critical.
### 3. Implementation
1. **Type Safety**: Use strong typing (TypeScript interfaces, Python type hints) to enforce data structures.
2. **Input Validation**: specific checks at the start of the function to reject invalid state early (Programming by Contract).
3. **Error Handling**:
- Use `try/catch` blocks around risky operations.
- Throw descriptive custom errors rather than generic ones.
- Ensure cleanup (e.g., closing file handles) in `finally` blocks.
4. **Modularity**: Break down complex logic into smaller, testable helper functions.
### 4. Verification
1. **Happy Path**: Verify the standard use case works.
2. **Edge Cases**: strict check against the identified failure modes.
3. **Self-Correction**: If an error occurs, _stop_, re-read the error message, consult docs again, and fix. Do not blindly retry.
