responsive-audit-inline-skill

In-session responsive audit delegate — wraps the playwright-responsive-audit-skill inline loop (detect → fix → re-verify) for current-diff frontend changes. Decision tree with skip rules, scope bounds

SKILL.md

GitHub

Responsive audit (inline)

You are executing the responsive-audit delegate's workflow in this session via the playwright-responsive-audit-skill inline loop — load that skill for the capture/fix tiers; this skill supplies the decision tree and discipline.

Decision tree

  1. Skip check (all must hold to skip):
    • No playwright.config.* + no @playwright/test in the project → report "Playwright not configured — skipping (never install unprompted)", stop.
    • The diff touches no frontend surface (components/**/*.{tsx,jsx,vue, svelte}, app|pages|routes|src/ui, rendered-route handlers) → report "no frontend surface in diff", stop.
  2. Load playwright-responsive-audit-skill and run its detect tier: mobile/tablet/desktop viewports against the diff's affected pages.
  3. Fix tier: auto-fixable defects (missing viewport meta, unbounded widths, overflow) fixed directly; layout-judgment defects reported, not guessed.
  4. Re-verify tier: re-capture only the fixed viewports/pages.

Scope bounds

  • Pages affected by the current diff only. No whole-site sweeps.
  • Fixes touch frontend files in the diff's blast radius; new dependencies are findings, never installs.

Enforcement deltas (vs responsive-audit-subagent)

Subagent enforcement Inline discipline (you)
Background execution + timeout model Runs block this session — state the capture plan first; abort and report if capture exceeds a reasonable interactive wait
Fresh context per audit Restate the affected-pages list before capturing so scope drift is visible
Tier model (vision) Same model as the caller — if you cannot assess the screenshots, report inconclusive instead of guessing

Output contract

Status: [success | partial | failed | skipped] Output: viewports × pages captured, defects found/fixed/reported (file:line), skip reason if any Summary: ≤3 sentences Issues: inconclusive captures, environment gaps, or "None"