# Review AIR Constraints for Soundness

AIR designs can contain subtle unconstrained states, selector gaps, and degree errors. This skill guides structured reviews that connect each protocol claim to enforceable constraints.

## Install

```bash
npx skillstore add clementwalter/air-cryptographer
```

## Metadata

- Status: approved
- Slug: clementwalter-air-cryptographer
- Skillstore revision: r2
- Version status: missing
- Tree hash: 9f826b7b890351b3000c40a7dfca9d90c9767cafbf158cf372abad3684a2f2c0
- Author: ClementWalter
- GitHub username: ClementWalter
- License: MIT
- Repository: https://github.com/ClementWalter/rookie-marketplace/tree/main/air-cryptographer/skills/air-expertise
- Ref: c68df504887c54bf71d2d467a1fbebde49467868
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands
- Quality score: 72
- Public page: https://skillstore.pages.dev/skills/clementwalter-air-cryptographer
- Manifest: https://skillstore.pages.dev/api/skills/clementwalter-air-cryptographer/manifest

## Capabilities

- Explains finite-field and polynomial concepts used in AIR systems.
- Reviews trace columns, row semantics, selectors, and boundary constraints.
- Identifies common soundness failures in transition, lookup, permutation, and memory arguments.
- Creates a constraint map from semantic claims to enforcing equations.
- Performs degree-accounting checks for quotient and composition polynomials.
- Proposes adversarial witness tests for AIR review.

## Use Cases

- Audit a new AIR design: Map each execution rule to constraints and identify unconstrained columns or selector gaps before implementation.
- Prepare a security review: Build an adversarial test plan for boundary conditions, memory consistency, and challenge binding.
- Learn AIR soundness methods: Study practical checks for trace design, degree limits, and common STARK vulnerability patterns.

## Prompt Templates

### Explain an AIR concept

```
Explain why every AIR trace column needs constraints. Use a small conceptual example without implementation code.
```

### Check selector design

```
Review this selector design: [describe selectors and gated constraints]. Check Booleanity, coverage, exclusivity, and ghost modes.
```

### Build a constraint map

```
Create a constraint map for this AIR: [describe columns, row meanings, and rules]. List each claim, its enforcing constraint, and gaps.
```

### Conduct an adversarial AIR review

```
Audit this AIR specification for soundness: [provide full specification]. Test transitions, boundaries, lookups, memory, quotient degree, and Fiat-Shamir binding. Rank findings and suggest fixes.
```

## Limitations

- It provides guidance and does not execute or formally verify an AIR implementation.
- It cannot replace an independent cryptographic audit of production proving systems.
- Its conclusions depend on the complete constraints and protocol details you provide.
- It does not implement proving, verification, or polynomial commitment code.

## Best Practices

- Provide complete column definitions, constraints, domains, and transcript order before requesting a review.
- Test every claimed rule with a counterexample trace that attempts to bypass it.
- Track constraint degrees after selectors and boundary factors are applied.

## Anti Patterns

- Do not treat documented column meanings as constraints.
- Do not assume selector coverage without enforcing it algebraically.
- Do not reuse Fiat-Shamir challenges across independent arguments.

## Security Audit

- Audited at: 2026-07-18T10:42:23.611\+00:00
- Summary: All 15 static findings are false positives caused by Markdown backticks and cryptographic-review language. The two reviewed files are educational AIR documentation and contain no executable commands, reconnaissance instructions, prompt injection, or data-exfiltration intent.

## Stats

- Views: 221
- Downloads: 7
- Favorites: 0
- Popularity score: 0