/visual-lint — Baseline-free rendered-UI defect detection

Codex Compatibility

When running this skill in Codex, translate Claude Code-only primitives before acting: AskUserQuestion -> chat/request_user_input, TodoWrite -> update_plan, Task/TaskCreate/TeamCreate/SendMessage -> spawn_agent/send_input/wait_agent when available and allowed, and EnterPlanMode/ExitPlanMode -> a concise chat plan plus explicit approval. Resolve Read/Write/Bash/WebSearch/WebFetch to Codex file/shell/web tools, and map ~/.claude/... paths to ~/.agents/... or ~/.codex/... unless the task explicitly targets Claude Code. The Agent-launched web-designer adjudicator maps to a Codex sub-agent spawn; if unavailable, run the rubric inline (lightweight mode).

Cursor Compatibility

When running this skill in Cursor Agent, translate Claude Code-only primitives before acting: AskUserQuestion -> AskQuestion; TodoWrite -> Cursor TodoWrite or an equivalent checklist; Task/Agent sub-agent flows -> Cursor Task (subagents), parallel Tasks, or run_in_background when allowed; EnterPlanMode/ExitPlanMode -> Plan mode (SwitchMode / CreatePlan) plus explicit user approval. Resolve Read/Write/Bash/web/search/MCP via Cursor Composer or Agent equivalents. MCP names written as mcp__server__tool typically map to call_mcp_tool. Map ~/.claude/... to ~/.cursor/skills/ unless the task explicitly targets Claude Code.

Screenshot a running app, run a structured defect rubric over the rendered pixels, and produce a findings report with cited evidence. Fixing is out of scope — this skill reports. Hand the report to a session that owns the source if the user wants the issues fixed.

Why this exists