Formalize Math Paper — Security skill for Claude Code
Agent Skill for formalizing a mathematics paper in Lean 4 with Mathlib: prove it, audit it against the paper, and package it for Palomar.
How to install Formalize Math Paper
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open vltanh/formalize-math-paper and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Formalize Math Paper does
Agent Skill for formalizing a mathematics paper in Lean 4 with Mathlib: prove it, audit it against the paper, and package it for Palomar. Works with Claude Code, Codex, Gemini CLI, Antigravity, GitHub Copilot, Cursor and other SKILL.md agents.
Alternatives in Security
- Anthropic Cybersecurity Skills — 734+ structured cybersecurity skills for AI agents · MITRE ATT&CK mapped · agentskills.io open standard · Work 3.8k ★
- Audit Agent Sessions — Analyze all agent sessions from the last 3 days across Claude, Codex, and Gemini 238 ★
- Amazon Skills — Free AI agent skills for Amazon sellers— keyword research, competitor analysis, listing audit & more 127 ★
README
formalize-math-paper
An [Agent Skill](https://agentskills.io) for formalizing a mathematics research paper in Lean 4 with Mathlib, in any field of mathematics. It works with any agent that supports the `SKILL.md` format: Claude Code, OpenAI Codex, Gemini CLI, Google Antigravity, GitHub Copilot, Cursor, OpenCode, Amp and others.
The skill guides the work end to end:
- read the paper and record its results, constants, citations and suspected typos;
- set up a Lean project on current Mathlib, using the module system;
- state every result first, and check each statement against the paper's TeX source;
- prove everything the paper proves (Stage 1), then every result it cites (Stage 2), until the
project has no
sorryand noaxiom; - verify: the axioms of every declaration, and Comparator;
- clean up: remove unused hypotheses and other warnings;
- write
REPORT.md, an audit of the paper against the formalization (errors and gaps, missing and redundant hypotheses, how the paper uses each cited result), andREADME.md; - package the project for the Palomar registry
(Challenge/Solution,
comparator.json,formalization.yaml, preflight).
It also covers turning an existing, never-compiled Lean draft into a project that builds.
Install
With the [`skills`](https://github.com/vercel-labs/skills) installer, which detects your agents and links the skill into each of them:
npx skills add vltanh/formalize-math-paper -g # for your user; omit -g to install into the current project
The installer puts the skill in `~/.agents/skills/`, which most agents read, and links it into the directories of those that do not. Antigravity reads user-wide skills only from `~/.gemini/config/skills/`, so link it there too:
mkdir -p ~/.gemini/config/skills && ln -s ~/.agents/skills/formalize-math-paper ~/.gemini/config/skills/
Or copy the repository into your agent's skills directory by hand:
git clone https://gith
Related Skills
Auditpaper
Audit a complete scientific paper: references, methodology, results, discussion, future works. Validates all c
Agents Orchestrator Skill
Orchestra: a cross-vendor execute → audit → escalate loop for CLI coding agents (Claude Code, Codex, Antigravi
Polymarket Skills
Composable Agent Skills for Polymarket prediction market trading. Paper-trading-first, security-audited. Works
Hackz Huntkit
Offensive-security playbooks for AI coding agents — IDOR, 403 bypass, CORS, finding validation and dupe-checki
Securevector AI Threat Monitor
Security & Observability for AI Agents — audit every tool call, enforce allow/block policies, catch prompt inj
Raindrop Collection Governance
Agent Skills-compatible skill: govern a Raindrop.io library's collection structure via raindrop-mcp — audit, c
Related Agents
Paf Reviewer
Runs one phone-a-friend relay or review command in the background and reports back a receipt plus the findings
CLI Subagents Antigravity
Runs a task in Google Antigravity (agy, Gemini models) on the user's Google account (CLI Subagents; extra Goog
Gemini Delegate
Delegate broad, purely reading analysis over a lot of material to Gemini (Antigravity CLI) — condense large di