Solve Math Rigorously
Forces a counterexample search before trusting a proof -- caught sin(x)+cos(x) exceeding 1 right at x=0 where naive spot-checks look fine.
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 16, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · ae9d744
Ran the real bundled math_verify.py counterexample checker on the plausible-looking claim 'sin(x)+cos(x) <= 1 for all real x': it found a genuine violation at x=0.00785 (value 1.00782) that boundary-only spot-checks at x=0, pi/2, pi, 3pi/2 (all give exactly +-1) completely miss -- a concrete, reproducible catch of a false claim a careless baseline pass would rubber-stamp as true.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 9/10
- Docs & honesty 5/5
What Solve Math Rigorously does
A solver-first math skill: map the problem, pick a method deliberately, solve in inspectable steps, then verify with sanity/structural/independent/tool-assisted checks before presenting confidence. Ships local SymPy-backed scripts (math_verify.py, math_visualize.py, math_table.py, math_photo_helper.py) and 15 reference playbooks for photo intake, proofs, and method selection. Triggers on solving equations/systems, proving or disproving a claim, optimization, checking an answer, or reading math off a photo/PDF/whiteboard; explicitly de-prioritizes teaching/worksheet modes unless asked.
How to install Solve Math Rigorously
git clone https://github.com/googlarz/math-skill
mkdir -p ~/.claude/skills
cp -r math-skill ~/.claude/skills/solve-math-rigorously
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Solve Math Rigorously
-
/solve-math-rigorouslyForces a counterexample search before trusting a proof -- caught sin(x)+cos(x) exceeding 1 right at x=0 where naive spot-checks look fine.
It also activates on plain-language prompts like these:
-
Prove or disprove that sin(x) plus cos(x) is always at most 1 -
Check my optimization solution for errors before I trust it -
Solve this system of equations and verify it with an independent check
Frequently asked questions
- Is the Solve Math Rigorously skill free?
- Yes. The skill itself is free from googlarz/math-skill. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Solve Math Rigorously work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 16, 2026. Verdict: Tested · Works. Ran the real bundled math_verify.py counterexample checker on the plausible-looking claim 'sin(x)+cos(x) <= 1 for all real x': it found a genuine violation at x=0.00785 (value 1.00782) that boundary-only spot-checks at x=0, pi/2, pi, 3pi/2 (all give exactly +-1) completely miss -- a concrete, reproducible catch of a false claim a careless baseline pass would rubber-stamp as true.
- What is the Solve Math Rigorously SkillProof Score?
- 9.6/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 9/10, docs & honesty 5/5.
- How do I install Solve Math Rigorously?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Solve Math Rigorously with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.