autospec-playwright

name	autospec-playwright
description	Use when you want to run disciplined no-mock Playwright UI-test authoring (Stage 2A) for a project. Detects the .autospec/test.yml authoring blocks, invokes autospec-test Stage 2A, and prints the coverage report.

autospec-playwright — Disciplined Playwright Authoring Dispatcher

Thin dispatcher for the autospec-test Stage 2A disciplined-authoring pipeline. Reads .autospec/test.yml authoring and reset blocks, invokes Stage 2A, and prints the e2e coverage report from e2e/.autospec/coverage.json. Carries NO machinery of its own — all logic lives in autospec-test.

Self-update mode

Decide this purely from the request text the harness handed you. Do NOT shell out (no grep, sed, [[ =~ ]], command substitution, etc.) to test the user's free-form request — passing it through a shell is what historically tripped harness permission engines. Read the request, normalize it in your reasoning (collapse whitespace, trim, lowercase), and if the result is exactly update, this skill enters self-update mode and does NOT run the normal pipeline.

1. Detect harness by checking which install path exists for this skill:
   • Claude Code: ~/.claude/skills/autospec-playwright/SKILL.md
   • OpenCode: ~/.config/opencode/agent/autospec-playwright.md
   • Codex CLI: ~/.codex/prompts/autospec-playwright.md
2. Re-install from main by piping the canonical installer:

   curl -fsSL https://raw.githubusercontent.com/berlinguyinca/autospec/main/bootstrap.sh | bash -s -- --skill all --harness all --update
   

   Run this one-liner once; it refreshes all autospec skills across all harnesses.
3. Show the diff between the prior installed file(s) and the freshly fetched copy.
4. Stop. Do not enter any pipeline phase. Print the upgrade summary and return to the user.

If no install path is detected, print Self-update: no installed copy of autospec-playwright found; run install.sh first. and exit.

Model tier

Model tier: TIER_B for all dispatch work (thin dispatcher; no design decisions).

Required capabilities & harness adapter

This workflow assumes a small set of capabilities. Map each one to your harness's actual tool.

Capability	Claude Code	OpenCode	Codex CLI	Fallback if missing
Read-only codebase research	Agent (subagent_type=Explore)	task agent in read-only mode	shell grep	Do the search in-thread
Foreground delegation	Agent (subagent_type=general-purpose)	nested task agent	spawn nested CLI session	Do the work in-thread
Ask the user a question	AskUserQuestion	inline prompt	inline prompt	Ask in response and wait
Subagent model tier	Model tier: TIER_B for all dispatch work	top task model + medium reasoning	top GPT + reasoning_effort=medium	Honor per-phase tier mapping
Subagent dispatch policy	per AGENTS.md decision matrix	per AGENTS.md decision matrix	per AGENTS.md decision matrix	inline with main-session token cost

Harness detection

Detect your harness by checking available tools before any phase:

1. Claude Code — the Agent tool with a subagent_type parameter is available.

   • TIER_A = opus + ultrathink
   • TIER_B = sonnet

2. OpenCode — a task tool with model/tier configuration is available (no subagent_type).

   • TIER_A = top-tier task model + high reasoning
   • TIER_B = smaller-tier task model + medium reasoning

3. Codex CLI — neither Agent nor a configurable task tool is available.

   • TIER_A = current top GPT model + reasoning_effort=high
   • TIER_B = gpt-5.1-codex-spark + reasoning_effort=medium

Fallback rule: If TIER_B is not available (quota, capacity, or authorization failure), silently retry with TIER_A. Never ask the user.

Invocation

/autospec-playwright [--dry-run]

• --dry-run — print what would be done; do not invoke Stage 2A.

Dispatcher flow

1. Locate config. Read .autospec/test.yml via loadAuthoringConfig() from skills/autospec-test/scripts/authoring-config.mjs. If the file is missing or e2e.authoring.enabled is false, print:

   autospec-playwright: e2e.authoring.enabled is false (or .autospec/test.yml missing).
   Set e2e.authoring.enabled: true to enable Stage 2A authoring.
   

   and exit 0.

2. Invoke Stage 2A. Call autospec-test Stage 2A — disciplined authoring — passing the resolved spec_dir, helpers_dir, route_clusters, coverage_target_pct, and fanout_max from the authoring config. Stage 2A is owned by skills/autospec-test/scripts/stage2a-orchestrator.mjs; this dispatcher does NOT duplicate any of its logic.

3. Print coverage report. After Stage 2A completes, read e2e/.autospec/coverage.json (shape: {total, covered, pct}) and print:

   autospec-playwright coverage: covered=<covered>/<total> (<pct>%)
   

   If the file is absent, print autospec-playwright: coverage.json not found — Stage 2A may not have run.

4. Exit code. Exit 0 on success, 1 on Stage 2A failure.

Loaded-data dashboard evidence

For autospec-created dashboard E2E or visual tests, do not accept a fixture, screenshot, or baseline until the test proves the dashboard is past its loading shell and has rendered service-backed data. The authored test must include at least one loaded-data assertion against a stable KPI, table row, service field, or route-specific datum that comes from the installed fixture response rather than from placeholder UI chrome.

Reject any visual baseline or snapshot whose visible page still matches /loading|skeleton|empty/i, including loading spinners, skeleton cards, empty states, or placeholder-only dashboards. If such text or UI is still present, wait for the loaded-data assertion to pass or fix the fixture route before capturing the baseline.

When dashboard tests install network fixtures, register broad catch-all mocks first, then install route-specific dashboard fixtures after catch-all mocks so the loaded service response wins over fallback handlers. The acceptance proof must cite the loaded-data assertion and the fixture route that supplies it.

Out of scope

• Stage 2A machinery (owned by autospec-test issue #996).
• Reset-endpoint generation (owned by autospec-test issue #997).
• Loop classification (owned by autospec-test issue #998).
• Any Playwright spec file authoring directly — this dispatcher wires; it does not author.