Claude Skills by victoraguilarsantamariadev

Exploratory end-to-end QA with a team of agents that drive a real browser like human testers — they click through the whole app, create/fill/submit, screenshot and visually judge rendering + data correctness, capture trace/HAR/console/video evidence, adversarially verify each finding, learn from rejected findings, and codify confirmed bugs (and working flows) as real regression/smoke tests. Use when asked to "test the whole app", "explore for bugs", "QA this web app", "find anything broken on...

Second half of the qa-explore loop — pick up the issues a human marked as real bugs in the tracker (GitLab/GitHub), reproduce each one, write a regression test that asserts the correct behaviour, fix the code until that test goes green and the suite stays green, then open a merge request linked to the issue for human review. Use when asked to "fix the QA issues", "work the qa-explore backlog", "auto-fix the confirmed bugs", or "open MRs for the labelled issues". Never merges — a human always ...

Release gate / QA sign-off — the decision layer a senior QA team owns. It aggregates every signal the other skills produce (explore findings by severity+confidence, the Step-0 deterministic suite, access-control, a11y, and optional visual/perf) and applies a WRITTEN, consistent release rubric to emit a deterministic GO / NO-GO verdict with the exact blockers, the risk summary, and an auditable list of waivers. The verdict is computed by rules (not LLM vibes); an agent only writes the human-re...

Self-healing test suite — when the deterministic E2E/regression suite goes red, triage each failing test into a STALE/BRITTLE test (repair the selectors/waits, never the assertion) vs a REAL REGRESSION (leave it red — it's a bug). Repairs the brittle ones on a branch and opens one merge request; real regressions are filed as issues that flow into /qa-fix. Use when asked to "fix the failing tests", "heal the broken suite", "the selectors are stale", "stop the flaky tests", or "the CI suite is ...

Living user / configuration manual generated by an agent that actually operates your app — the same drive-your-app engine as qa-explore, pointed at a different output (docs instead of bugs). It recons the app, ORDERS the features into a real setup/usage sequence, proposes a table of contents + audience for you to approve, then walks the happy path in that order, screenshots and annotates every step, and assembles one coherent Markdown manual. The Markdown is the master, so it re-generates whe...

Risk-based test plan — what a senior QA team writes BEFORE testing. Recon the app, score each area by RISK (impact × likelihood), rank into P0/P1/P2 deterministically, define the acceptance ("done") per area, and emit a test charter (Markdown) that also seeds qa-explore's areas in priority order so the expensive exploration attacks the riskiest surface first. Use when asked to "write a test plan", "what should we test", "risk assessment", "test strategy", "prioritize testing", or "QA plan for...