---
title: "math-proof: siege: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/skills/anthropic-math-proof-siege/
description: "Runs an hours-long, multi-round attack on one hard mathematics problem with judge and worker sub-agents and a bookkeeping script, ending in a proof.md whose Status section says what is and is not proved."
---

# math-proof: siege: review, setup & limitations · Undominated.ai

> Runs an hours-long, multi-round attack on one hard mathematics problem with judge and worker sub-agents and a bookkeeping script, ending in a proof.md whose Status section says what is and is not proved.

[← Explore all skills](/skills/)

RESEARCH AND DOCUMENTATION / Anthropic

# math-proof: siege

Runs an hours-long, multi-round attack on one hard mathematics problem with judge and worker sub-agents and a bookkeeping script, ending in a proof.md whose Status section says what is and is not proved.

 [See setup guidance ↓](#setup)[Original source ↗](https://github.com/anthropics/claude-plugins-official/tree/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege)

SOURCE REVIEW

 Reviewed 2026-10-22
 Evidence 7 linked sources
 Publisher Anthropic
 Licence [Apache-2.0 ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/LICENSE)
 Revision 41f32e407b45
 [Read what was—and wasn’t—checked ↓](#review)

“Work on one hard mathematics problem in rounds: each round a judge writes a few self-contained questions, independent workers answer them, and the judge keeps a ledger of what is proved, refuted and open, until the judge concludes or the rounds run out, and then a proof.md says plainly what is and is not proved. A run takes hours and dozens of worker runs. Usage: /math-proof:siege [NAME=value settings] <the problem, stated in full, or the path of a file holding it>.”

 [Anthropic · upstream description ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege/SKILL.md) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - The orchestrating session does no mathematics: each step goes to a fresh math-proof-judge subagent and each query to a fresh worker that sees only that query and the problem, and the judge keeps a ledger of claims marked PROVED, REFUTED or OPEN.
- Mechanical decisions are made by scripts/ledger.py, a standard-library Python script that validates the judge's output, applies the early-conclusion gate and decides which worker answers are finished; a complete proof or disproof of the goal is checked line by line by two more workers before the judge concludes.

### A good fit for

 - Run a long multi-round attempt on a research-level problem and get a proof.md that separates what is proved from what is open
- Resume a stopped run by giving the same /math-proof:siege line again in the same directory
- Shorten a run or change its wave sizes with settings such as MAX_ROUNDS=8 or WAVE_DEEP=6 placed before the problem

### Weigh up before choosing

 - Last commit touching plugins/math-proof/skills/siege at revision 41f32e40: 2026-09-30 (aa5654b7).
- The judge runs model-written Python through Claude Code's shell tool on the user's machine. The README says to use --dangerously-skip-permissions only in a disposable container or VM, as a non-root user, and otherwise to approve its commands by hand.
- The README says to expect dozens of sub-agent runs, and well over a hundred if all fourteen rounds run, and to check /usage on a subscription or cap spend with --max-budget-usd on an API key.
- Manual invocation only: disable-model-invocation is true.
- proof.md is the model's own account and literature results are cited from memory, per the README.
- Its settings block keeps sub-agents in the foreground by turning background tasks off and gives a silent worker four hours, for every session that reads that settings file until it is removed.

02 / THE REVIEW RECORD

## What we actually inspected.

Source review has boundaries. A clear record is more useful than a “safe” badge.

### Material inspected

 - plugins/math-proof/skills/siege/SKILL.md (frontmatter, approach, arguments and settings)
- plugins/math-proof/skills/siege/scripts/ledger.py (module docstring and imports)
- plugins/math-proof/agents/math-proof-judge.md
- plugins/math-proof/agents/math-proof-worker.md
- plugins/math-proof/agents/math-proof-worker-deep.md (frontmatter and tool sentence)
- plugins/math-proof/README.md
- plugins/math-proof/LICENSE

### Our findings

 - Default settings: ESC_ROUND 4, WAVE 4, WAVE_DEEP 10, MAX_ROUNDS 14, MIN_ROUNDS 4, REFINE_STEPS 2 and MAX_COMMIT 9; DIR=path is also accepted, and any other NAME=value token (two or more capitals, a whole-number value) at the start of the message is refused.
- From round ESC_ROUND on, every worker is a math-proof-worker-deep and a round may hold up to WAVE_DEEP queries.
- Worker agents have Read, Write and Edit only and are told they have no web access and no code execution; the judge agent has Read, Write, Edit, Bash, Glob and Grep and may run Python 3 checks.
- ledger.py imports only re, sys and pathlib (besides from __future__ import annotations), needs Python 3.7 or later, and prints a fail-closed verdict instead of a traceback on internal errors.

### Not established by this review

 - No siege run was started; ledger.py was not executed.
- The SKILL.md was read through its settings section and grepped for state.md, result-so-far.md and the Status section; the round loop and proof-tail steps were not read line by line.
- Not executed: it was not installed or run.

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

## Add a skill to your workflow.

[Upstream setup instructions ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/README.md)
 - The math-proof README's install line, `/plugin install math-proof@claude-plugins-official`, adds the whole math-proof plugin (siege, solo and the three agents siege launches), so it is not siege's own command and install.command is null.
- Before the first run the README asks for the "env" block shown in configExample, in ~/.claude/settings.json or in .claude/settings.json of the directory you run from.
- The README also prints an unattended form, claude -p "/math-proof:siege problem.md" with --model claude-opus-5-5, --effort high and --dangerously-skip-permissions, and says adding --max-budget-usd <amount> stops the run once that much has been spent on an API key.
- No line was run for this review.

### Before you start

 - Claude Code 2.1.280 or later
- Python 3.7 or later on PATH as python3 or python, with no extra packages

### Documented configuration example

Use the client and authentication context described in the setup guidance. Configuration formats vary between hosts.

 {
 "env": {
 "CLAUDE_CODE_MAX_OUTPUT_TOKENS": "128000",
 "CLAUDE_CODE_DISABLE_BACKGROUND_TASKS": "1",
 "CLAUDE_ASYNC_AGENT_STALL_TIMEOUT_MS": "14400000"
 }
} Copy configuration ↗

### Compatibility

Claude Code

### Spawns judge and worker subagents and runs a local Python bookkeeping script; the judge has an unrestricted shell; no web tools

 - allowed-tools: Read, Write, Edit, Glob, Grep, Agent, python3 or python running scripts/ledger.py, and Bash(mkdir, cp, mv, cat, test, ls, wc, cmp, grep, printf, echo)
- disallowed-tools for the orchestrating session: WebSearch, WebFetch, AskUserQuestion
- The math-proof-judge agent's tools include Bash with no command limit; its instructions use the shell for Python checks, splicing text with cat, sed or heredocs, and plain file handling, and the README says Claude Code asks the user to approve those commands in the default permission mode. Worker agents have Read, Write and Edit only
- Writes the run folder, ./math-proof-run by default, including state.md, ledger.md, query and answer files, proof.md and result-so-far.md

### Cost model

Apache-2.0 per plugins/math-proof/LICENSE; dozens to over a hundred sub-agent runs per problem, billed to the user's subscription or API key

04 / FOLLOW THE EVIDENCE

## The source trail.

Our notes are separate from the original resource. Check upstream before adopting a new version.

 - [math-proof siege SKILL.md ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege/SKILL.md) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege/SKILL.md Supports: upstreamDescription, summary, whySelected[0], whySelected[1], bestFor[0], bestFor[1], bestFor[2], limitations[3], review.findings[0], review.findings[1], access.permissions[0], access.permissions[1], access.permissions[3]
- [math-proof siege scripts/ledger.py ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege/scripts/ledger.py) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/skills/siege/scripts/ledger.py Supports: whySelected[1], review.findings[3], install.prerequisites[1]
- [math-proof-judge agent ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-judge.md) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-judge.md Supports: review.findings[2], access.permissions[2], limitations[1]
- [math-proof-worker agent ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-worker.md) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-worker.md Supports: review.findings[2], access.permissions[2]
- [math-proof-worker-deep agent ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-worker-deep.md) Checked 2026-10-22 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/agents/math-proof-worker-deep.md Supports: review.findings[2], access.permissions[2], review.findings[1]
- [math-proof README ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/README.md) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/README.md Supports: install.sourceUrl, install.steps[0], install.steps[1], install.steps[2], install.prerequisites[0], install.prerequisites[1], configExample, limitations[1], limitations[2], limitations[4], limitations[5], access.cost, compatibility, bestFor[1], bestFor[2], summary
- [math-proof LICENSE ↗](https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/LICENSE) Checked 2026-10-21 https://raw.githubusercontent.com/anthropics/claude-plugins-official/41f32e407b450055044f167db7ded5e8ff67d203/plugins/math-proof/LICENSE Supports: license, access.cost

KEEP COMPARING

## Other approaches to consider.

Related by category or shared topics. These are alternatives to inspect, not a measured quality order.

 [### math-proof: solo ↗ Works one hard mathematics problem in the current Claude Code session, writing each settled step to a notes file as it goes and ending with a self-contained proof.md.](/skills/anthropic-math-proof-solo/)[### Anthropic Documentation Coauthoring ↗ Guides a human and agent through audience discovery, section-by-section drafting and a separate-reader review of a substantial document.](/skills/anthropics-doc-coauthoring/)[### Claude Academy guide ↗ Answers a Claude how-to first, then may name at most two Academy items copied from a fetched catalog.](/skills/anthropics-academy-guide/)

 [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/)

## Continue your investigation

 - [Choose a toolkit](/tools/)
- [Inspect execution hosts](/agents/)
- [Check tool access](/mcp-servers/)
- [Check project dependencies](/check-repo/)
