bug-reproduce
Turn a known bug into a tight, red-capable reproducer, then prove the reproducer locks that…
Perform scoped exploratory QA of a web application, reproduce defects, collect screenshots and console evidence, and deliver a prioritized report.
$ npx -y skills add Prismer-AI/PrismerCloud --skill webapp-qa --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/webapp-qaContext preview
The summary Claude sees to decide when to auto-load this skill.
Perform scoped exploratory QA of a web application, reproduce defects, collect screenshots and console evidence, and deliver a prioritized report.
name: webapp-qa scope: common description: Perform scoped exploratory QA of a web application, reproduce defects, collect screenshots and console evidence, and deliver a prioritized report. license: MIT metadata: nativeReplaces: [dogfood] upstream: Hermes Agent skills/software-development/dogfood at 1a1f4a59e2
For the full five-phase exploratory method, read references/dogfood/GUIDE.md; its taxonomy and report-template links resolve within this bundle.
Use the user-supplied target URL, scope and existing test accounts. Inspect the available browser tools or Playwright setup first; never invent browser tool names. Without a working browser, report the blocker and limit claims to checks actually performed.
1. Map the main user journeys, navigation, forms and error/empty states. 2. Exercise each journey with positive and negative inputs. Inspect console and network failures after navigation and significant interactions. 3. Inspect screenshots at desktop and mobile sizes for clipping, overlap, loading failures and keyboard accessibility. 4. Reproduce each suspected defect. Record URL, exact steps, expected/actual behavior, console evidence and screenshot file. Separate application bugs from unavailable dependencies or test-environment failures. 5. Deduplicate, prioritize by user impact, and use references/issue-taxonomy.md plus templates/dogfood-report-template.md as report structure. 6. State tested, untested and blocked areas. Passing HTTP status alone is not a passing user journey. Never fill missing evidence with plausible descriptions. 7. Attach the report and requested evidence as task-bound assets using the actual platform delivery contract. Do not invent flags or use MEDIA markers.
Use a task-owned local fixture for browser-tool smoke tests before touching a live account. Record desktop/mobile viewport, keyboard, console, clipping and positive/negative form results independently; a fixture does not certify the user's real application.
Keep test mutations within the user's authorized scope. Do not submit real payments or send external messages as incidental QA actions.
Repo: Prismer-AI/PrismerCloud
Turn a known bug into a tight, red-capable reproducer, then prove the reproducer locks that…
Review a diff against its acceptance criteria in four segments (convention adherence, bug…
Five-dimension design audit (frontend UI/UX · server data-model & flow · endpoint spec ·…
Before merge, mechanize Documentation-First — derive the code delta from git diff, then…
Diagnose the local dev machine before any APC loop step — run apc env doctor, classify each…
Close out a local coding task on the bound daemon — stage, commit, branch, merge, push via…