BenSheridanEdwards

Audit a TypeScript/React frontend against an opinionated WCAG 2.2 AA baseline (tooling, component patterns, application shell) with optional implementation plan.

Audit a project's agentic instruction files (CLAUDE.md, AGENTS.md, .cursor/rules, .cursorrules, .github/copilot-instructions.md, Windsurf, Aider) and Claude Code settings (.claude/settings.json, .claude/settings.local.json) against an opinionated baseline covering project context coverage, operational guidance, settings hygiene, and multi-agent drift. Static-only with optional implementation plan.

Graphify-powered architectural audit of a TypeScript codebase. Produces a diagnostic snapshot + checks module boundaries, coupling, state/data flow, and conventions, with optional implementation plan.

Audit build pipeline and bundle output (static-first with optional --with-stats enrichment). Checks configuration, composition, hygiene, and performance with optional implementation plan.

Audit Node.js dependency tree (security, health, compliance, hygiene). Static-first with optional --with-network enrichment for vulnerabilities, outdated, and abandonment data. Optionally generates an implementation plan.

Audit project documentation against an opinionated baseline spanning onboarding, architectural/decision docs, code-level docs, and operational documentation with drift detection. Static-first with optional --with-link-check. Optionally generates an implementation plan for the gaps.

Audit a TypeScript codebase's error-handling discipline against an opinionated baseline spanning throw and catch hygiene, async and network error handling, React error boundaries, and logging and observability. Static-only by design. Optionally generates an implementation plan for the gaps.

Audit a TypeScript project's lint configuration against an opinionated baseline spanning configuration shape, rule coverage, strictness and enforcement, and suppressions hygiene. Auto-detects ESLint or Biome. Static-first with optional --with-run enrichment. Optionally generates an implementation plan for the gaps.

Use when a change touches device or Appium behaviour (driver, gestures, locators, ios/adb helpers, runner) and the Definition of Done requires running the focused device lane, or when a PR needs on-device screenshot proof. Explains how to run a real spec on a booted simulator/emulator and where the evidence lands.

Use when cutting a NativeProof release, bumping the package version, or writing a CHANGELOG entry. Covers the Keep a Changelog entry format and the GitHub-Release-driven npm publish flow.

Use when a change touches the CLI, init/onboard scaffolding, or the generated project surface (src/cli.ts, templates, runner-config) and the Definition of Done requires proof that the generated project still works. Produces that proof from the packed tarball, not from the repo checkout.

Use when opening or updating a GitHub pull request that needs screenshots, videos, or proof artifacts rendered inline in the PR description instead of shown as plain links.

Create and update Linear issues for OpenClaw. Use when the user asks to create a Linear ticket, mentions OpenClaw's Linear skill, asks to file approved suggestions in Linear, or needs help with the project's Linear workflow.

Use the OpenClaw browser tool for research, form-filling, and UI automation. Use when web_fetch is insufficient (JS-heavy pages, multi-tab research, authenticated flows).

Read and control I2C and SPI peripherals on Sipeed boards (LicheeRV Nano, MaixCAM, NanoKVM).