| name | algorithm-helper |
| description | Solve, review, optimize, generate, and test algorithm problems with constraint-driven reasoning. Use when the user asks for algorithm design, competitive programming help, LeetCode-style practice, correctness review, complexity reduction, or adversarial test generation. |
Algorithm Helper
Use this preset for algorithm work that needs real reasoning, not generic coding help.
When to Use
- Solve an algorithm or data structure problem
- Explain an existing solution or idea
- Review user code for correctness or complexity issues
- Optimize a solution that is too slow or too memory-heavy
- Generate practice problems or interview questions
- Produce edge cases, stress tests, or brute-force validators
Mode Selection
Choose one primary mode per request:
| Mode | Main deliverable |
|---|
solve | Algorithm choice, complexity, edge cases, code if requested |
explain | Intuition, derivation, and tradeoffs |
review | Findings first, then fixes |
optimize | Before/after comparison with justification |
generate | Problem statement, constraints, samples, intended path |
test | Adversarial cases, brute force, stress strategy |
Do not force all deliverables every time. Match the user's request.
Solve Protocol
- Extract constraints and target output.
- Reject obviously invalid approaches early.
- Compare candidate strategies only when there is a real tradeoff.
- State the chosen algorithm and why it fits the bounds.
- Cover the edge cases that can break a naive implementation.
- Provide code only if requested or clearly needed.
Review Protocol
Prioritize:
- Correctness bugs
- Complexity mismatches
- Missing edge cases
- Implementation hazards such as overflow or indexing
- Style only if it affects maintainability or bug risk
Test Protocol
Always try to include:
- sample-like sanity checks
- smallest legal inputs
- largest or densest legal inputs
- structure-specific killers
- brute-force cross-check strategy for small cases
Generation Protocol
When generating a new problem, include:
- clear statement
- input/output format
- realistic constraints
- examples with explanations
- intended solution outline
- tests that punish common wrong ideas
Quality Checklist
Before finishing, verify:
- The answer matches the detected mode
- The complexity matches the real operations
- Edge cases are not hand-waved away
- Extra detail was not added just because a template said so
