vnmoorthy

Fast headless browser for QA testing and site dogfooding. Navigate any URL, interact with elements, verify page state, diff before/after actions, take annotated screenshots, check responsive layouts, test forms and uploads, handle dialogs, and assert element states. ~100ms per command. Use when you need to test a feature, verify a deployment, dogfood a user flow, or file a bug with evidence. Use when asked to "open in browser", "test the site", "take a screenshot", or "dogfood this". (gstack)

Fast headless browser for QA testing and site dogfooding. Navigate pages, interact with elements, verify state, diff before/after, take annotated screenshots, test responsive layouts, forms, uploads, dialogs, and capture bug evidence. Use when asked to open or test a site, verify a deployment, dogfood a user flow, or file a bug with screenshots. (gstack)

A completion-claim gate. When active, the agent cannot assert that code work is done, complete, fixed, ready, shipped, implemented, or working unless the same turn also contains machine-checkable verification evidence. Applies to every coding task in the session. Not optional. Load this skill whenever you are about to write, edit, or modify code the user will run.

Use this skill whenever the user asks about their Claude Code spend, cost, token usage, cache efficiency, sessions, or anything related to "what is Claude costing me". Triggers include "how much did I spend", "why did my bill go up", "what session was most expensive", "do I have cache problems", "show me my Claude Code usage", "any sessions waste a lot of cache", "spend by project", "spend this week / month / today". This skill calls the local `ccmeter` CLI which reads `~/.claude/projects/*.jsonl` with no network access, then narrates the result. Do NOT trigger for questions about API pricing in the abstract or unrelated billing topics — this is specifically for the user's own Claude Code session-log analysis.

Run after finishing any coding task to catch the 35 specific failure modes AI agents commit (dead code, hallucinated APIs, drive-by refactors, silent catches, spec drift). Use this skill when the user is wrapping up a code change, just before declaring "done", or whenever they want a review of the latest diff. Triggers on phrases like "review my changes", "before I commit", "check what I just did", "lint my diff", "is this clean", or any moment you've just made code edits across files.