mathguard - Math-heavy optimization for large-scale algorithms
Guides exact or bounded-approximation algorithm choices for large-scale, math-heavy workloads with explicit errors, bounds, and trade-offs.
Tags
Updated: 2026-10-04Capabilities
What this skill does
- Assess classical algorithmic floors
- Select named math techniques
- Declare exact or approximate modes
- Specify ε/δ error parameters
- Explain asymptotic bounds
- State optimization trade-offs
- Reject unsafe approximations
- Provide code or pseudocode
Inputs
- Problem scale
- Workload characteristics
- Classical algorithm and bound
- Exactness requirements
- Caller error tolerance
- Performance bottleneck
- Team familiarity
Outputs
- Algorithm recommendation
- Exactness and error declaration
- Asymptotic bound justification
- Trade-off statement
- Code or pseudocode
Requirements
- Compatible coding agent: Claude Code, Antigravity, Cursor, Gemini CLI, or Codex CLI
- Prior lemmaly assessment
