michaelalber/rust-architecture-checklist

Rust Architecture Checklist

"A checklist cannot fly a plane, but a pilot cannot fly safely without one." — Atul Gawande

Core Philosophy

Shared across the dotnet / python / php / rust architecture checklists — same values, language-specific checks.

| # | Value | What it means | |---|-------|---------------| | 1 | Detect before judge | Determine version/framework/structure before applying any item; context decides what is idiomatic. | | 2 | Evidence over opinion | Every finding cites file:line and the offending pattern. "Overuses clone" is not a finding; "handlers/order.rs:47 clones a Vec<OrderItem> that could be borrowed" is. | | 3 | Feature cohesion | Organized by business capability, not technical layer. Cross-feature coupling is a violation. | | 4 | Dependencies point inward | Domain logic does not depend on I/O or framework crates. Boundaries are explicit (pub(crate) by default). | | 5 | Explicit error handling | Failures handled at the right layer; no silent swallowing; errors carry diagnostic context. | | 6 | Config & secrets hygiene | No hardcoded secrets; configuration injected, not reached for globally; secrets from env / secret-manager. | | 7 | Version awareness | Recommendations are gated to the detected edition; never suggest an idiom that does not exist there. | | 8 | Tests gate change | Untested code is a finding; high-risk modules without tests are prioritized. | | 9 | Graded, actionable output | A letter grade (A–F) from counted findings, plus prioritized, edition-correct recommendations. |

Workflow

Shared skeleton: DETECT → SCAN → REPORT → RECOMMEND.

DETECT     Rust edition (Cargo.toml `edition`), async runtime (Tokio/async-std/none), and structure
           (workspace vs single crate). Record findings; edition changes what is idiomatic.

SCAN       Run the Rust Checklist below section by section. Gather evidence with tooling:
             cargo clippy -- -D warnings   # baseline — must pass before the checklist
             cargo audit                   # CVEs in the dependency tree
             grep -rn "unsafe" src/        # enumerate unsafe blocks for the audit
           Every violation becomes a finding with file:line and a severity (critical/high/medium/low).

REPORT     Emit the graded report (Output Template). Grade = function of counted findings.

RECOMMEND  Prioritize: critical → quick wins → modernization. Edition-gate every recommendation.

Rust Checklist (language-specific)

| # | Check | Severity | |---|-------|----------| | 1 | Ownership discipline — no .clone()/.to_owned() where a borrow suffices; Rc/Arc only for genuine shared ownership | High | | 2 | Trait coherence — trait impls have clear semantic purpose; no "marker-trait soup" to satisfy generic bounds | Medium | | 3 | Error handling — thiserror for library errors, anyhow for application; no .unwrap()/.expect() outside #[cfg(test)]; consistent ? | High | | 4 | unsafe justification — every unsafe block carries a correct // SAFETY: comment; blocks without one are automatic High findings | Critical | | 5 | Crate boundary hygiene — pub(crate) is the default; pub is a deliberate API decision; internal types do not leak | High | | 6 | Async runtime consistency — exactly one runtime per binary; no blocking calls (std::thread::sleep, sync I/O) inside async fn | High | | 7 | Dependency discipline — cargo audit clean; cargo deny for licenses/duplicates; minimal version constraints | Medium | | 8 | Test organization — unit tests in #[cfg(test)], integration tests in tests/; behavioral, not structure-coupled | High |

Clippy config: clippy configuration. Full section-by-section list (with the unsafe audit table): review checklist.

State Block

<arch-checklist-state>
language: rust
mode: DETECT | SCAN | REPORT | RECOMMEND | COMPLETE
detected: [edition | runtime | workspace/single | tests:yes/no]
issues_found: [critical:N high:N medium:N low:N]
last_action: [what was just done]
next_action: [what should happen next]
</arch-checklist-state>

Output Template

Shared across all four architecture checklists.

## Architecture Checklist: [crate] (Rust)
**Edition**: [2015/2018/2021/2024] | **Runtime**: [Tokio/async-std/none] | **Tests**: [yes/no]

| Section | Pass | Fail | Warn |
|---------|------|------|------|
| Ownership / Traits / Errors / unsafe / Boundaries / Async / Tests | … | … | … |

### Grade: [A–F]
Grading: **A** 0 crit/0 high/≤3 med · **B** 0 crit/≤2 high · **C** 0 crit, gaps in one area ·
**D** 1+ crit · **F** fundamental problems (unjustified `unsafe`, runtime conflict, library `.unwrap()` everywhere).

| Severity | Location | Finding | Recommendation |
|----------|----------|---------|----------------|
| CRITICAL | file:line | [pattern] | [edition-gated fix] |

**`unsafe` audit**: | location | SAFETY comment | quality | risk |
**Quick wins**: [low-effort, high-impact] · **Modernization**: [larger items with effort estimate]

AI Discipline Rules

• Clippy must pass first. cargo clippy -- -D warnings is the baseline; report failures before the architectural checklist.
• Evidence or it is not a finding. Cite file:line; show the grep/clippy output. Never grade on vibes.
• Edition-gate recommendations. Do not flag 2015-era patterns in a 2015 crate, or suggest 2024-edition idioms (e.g. RPIT lifetime capture, unsafe-attribute syntax) for an older edition.
• Architecture, not security. Memory-safety-as-vulnerability and crypto findings belong to rust-security-review — note them and route there.

Integration with Other Skills

• architecture-review — When the grade is D/F, escalate to the Socratic critic: this checklist finds what is wrong; architecture-review builds why.
• rust-security-review — Companion for the security dimension (unsafe audit for memory-safety vulns, cargo-audit).
• rust-feature-slice — Correct-pattern reference when the checklist flags structural violations.
• tdd — Methodology for adding tests the checklist flags as missing, and for driving any refactor.
• dotnet / python / php-architecture-checklist — Sibling skills sharing this exact Core Values + workflow + output.