skillsSource-backedReview first Safety ✓ Privacy ✓
Playwright Trace Failure Triage Capability Pack Skill
Expert skill for reviewing Playwright trace artifacts, screenshots, action timelines, network events, retries, and CI evidence to classify flaky browser test failures without guessing from logs alone.
by JSONbored·added 2026-06-05·
Claude CodeCodexWindsurfGeminiCursorCLI
HarnessClaude CodeCodexWindsurfGeminiCursorCLI
Level:expertType:capability-packVerified:validated
Review first — review before installing
Open the source and read safety notes before installing.
Safety notes
- Trace artifacts can include screenshots, DOM text, URLs, request metadata, console output, and application state; review them before sharing publicly.
- Do not fix a trace-only symptom by adding broad waits, retries, or timeouts unless the trace evidence supports that change.
- Keep destructive browser actions and production-like credentials out of replayed failure reproduction.
Privacy notes
- Playwright traces can expose user names, emails, tokens in URLs, internal hostnames, test data, screenshots, and API payload fragments.
- Public PR comments should summarize trace evidence without uploading private trace files or pasting sensitive network details.
Prerequisites
- Playwright test failure, trace artifact, test report, or CI run under review.
- Access to test source, Playwright config, browser/project name, retry settings, and relevant artifacts.
- Permission to inspect screenshots, DOM snapshots, console logs, network events, and request/response metadata.
Schema details
- Install type
- package
- Reading time
- 9 min
- Troubleshooting
- No
Source repository stats
- Scope
- Source repo
Package metadata
Skill and platform metadata
- Skill type
- capability-pack
- Skill level
- expert
- Verification
- validated
- Verified at
- 2026-06-05
Retrieval sources
https://playwright.dev/docs/trace-viewer-introhttps://playwright.dev/docs/debughttps://playwright.dev/docs/test-configurationhttps://playwright.dev/docs/test-reportershttps://github.com/microsoft/playwright
Tested platforms
ClaudeCodexCursorWindsurfGeneric AGENTS
| Platform | Support | Install path |
|---|---|---|
| claude-code | Native | .claude/skills/<skill-name>/SKILL.md |
| codex | Native | .agents/skills/<skill-name>/SKILL.md |
| windsurf | Native | .windsurf/skills/<skill-name>/SKILL.md |
| gemini | Native | .gemini/skills/<skill-name>/SKILL.md or .agents/skills/<skill-name>/SKILL.md |
| cursor | Adapter | .cursor/rules/<skill-name>.mdc |
| cli | Manual | AGENTS.md or tool-specific context file |
Full copyable content
# Trigger
"Apply the Playwright trace failure triage capability pack to this failed test."
# Required output
1) Artifact inventory and Playwright version evidence
2) Failure classification from trace actions, DOM snapshots, network, console, and retries
3) Root-cause hypothesis with supporting trace evidence
4) Smallest safe fix and regression validation planAbout this resource
Knowledge Freshness
This capability pack is grounded in the current Playwright Trace Viewer, debugging, test configuration, and reporter documentation checked on 2026-06-05. Prefer the official Playwright docs and repository over remembered behavior when reviewing trace retention, report fields, or CLI commands.
Retrieval Sources
- https://playwright.dev/docs/trace-viewer-intro
- https://playwright.dev/docs/debug
- https://playwright.dev/docs/test-configuration
- https://playwright.dev/docs/test-reporters
- https://github.com/microsoft/playwright
Core Workflow
- Identify the failed test, browser project, Playwright version, retry count, trace retention setting, and CI environment.
- Inventory artifacts: trace, report, screenshots, video, console output, network events, test stdout, and any attached application logs.
- Open the trace timeline and find the first user-visible failure, not only the final assertion message.
- Compare action snapshots before and after the failing step. Classify whether the failure is locator ambiguity, timing, navigation, network, auth, data, accessibility role mismatch, animation, browser-specific behavior, or product regression.
- Check retries. A pass-on-retry usually needs flake classification and a deterministic reproduction attempt, not an automatic merge.
- Recommend the smallest fix: stable locator, fixture setup, route mock, assertion change, app bug fix, or test isolation change.
- Provide a validation plan with the exact Playwright command, browser project, retry setting, and artifact retention expected after the fix.
Capability Scope
- Trace viewer evidence review
- Timeline/action classification
- DOM snapshot and screenshot comparison
- Console and network event triage
- Retry and CI artifact interpretation
- Locator, assertion, fixture, and isolation recommendations
- Privacy-safe public summary of trace evidence
Compatibility
Native
- Claude Code and Claude: use as a reusable skill for Playwright failure review.
- Codex workflows: use for PR review and CI failure triage where artifacts are available.
Manual Adaptation
- Generic AGENTS files: convert the workflow into a browser-test failure review checklist.
- Cursor and Windsurf: use the output contract when reviewing trace artifacts from CI.
Required Inputs
- Failed test name, file, browser project, and retry number
- Playwright config and trace retention setting
- Trace artifact or report link
- Relevant test source and fixture setup
- Expected product behavior and environment details
Production Rules
- Do not approve a change only because the test passed on retry.
- Do not suggest arbitrary waits without trace evidence that the application was still legitimately loading.
- Do not expose screenshots, URLs, request payloads, or DOM text from private traces in public comments.
- Prefer role-based or user-visible locator fixes when the trace shows selector brittleness.
- Keep app regressions separate from test bugs; a failing trace can prove the product is wrong.
- Require a rerun of the affected browser project after the fix.
Output Contract
- Artifact inventory and confidence in the available evidence.
- Failure classification with trace-backed observations.
- Root cause hypothesis and alternatives ruled out.
- Minimal fix recommendation.
- Validation commands and artifact expectations.
- Privacy notes for anything that should not be pasted publicly.
#playwright#trace-viewer#testing#browser-automation#flaky-tests
Source citations
Signals
Loading live community signals…
More like this, weekly
A short, calm digest of reviewed Claude resources. Unsubscribe any time.