moonshineaitech/lunastack/perf-budget
Claude Code and agent skill by moonshineaitech/lunastack
- community
Checking that the SKILL.md file is still in the repository.
Use when starting a user-facing feature or reviewing a PR that adds JS/CSS/images/fonts or new API endpoints. Set numeric performance budgets BEFORE building and enforce them in CI, so weight and latency constrain decisions instead of being discovered after ship.
Install
npx vskill@latest install moonshineaitech/lunastack/perf-budgetSKILL.md
The same SKILL.md was published earlier as moonshineaitech/lunastack/perf-budget.
/perf-budget — Performance Budget
Role: Performance Lead. Set budgets BEFORE building, enforce during.
Enforcement rule: block the PR if ANY single budget line is exceeded -- no averaging across pages, no "close enough." A 205KB JS bundle fails the 200KB budget, period. Even when a metric is still under budget, flag it if it regressed more than 10% from the last green build -- creep compounds. Cap "we'll fix it later" exceptions at zero: budget failures are merge blockers, not warnings.
Every number here must come from a real measurement -- Lighthouse or WebPageTest for load metrics, the bundle analyzer for weight, RUM or a load test for API percentiles. If a value wasn't measured, write "not measured" -- never estimate, back-solve from a target, or invent it.
BAD: "Ship it, we'll optimize later -- Lighthouse scored 92 on my laptop." (lab number on fast hardware, no budget, no enforcement) GOOD: "PR blocked: JS bundle measured 240KB > 200KB budget. Code-split the dashboard charting lib (55KB) behind a dynamic import, re-measure before merge."
Skip when: internal tools, admin dashboards, or throwaway prototypes with no external users and no mobile traffic -- budgets add friction with no bounce-rate payoff there.
Gotchas: Don't set budgets after building -- set them before, so they constrain decisions during development. Don't measure only lab performance (Lighthouse) -- Real User Monitoring shows what actual users experience on real devices. Don't let JavaScript budget creep past 200KB compressed -- every KB beyond that measurably increases bounce rate on mobile.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingperf-budget/SKILL.md
Popularity
- Repository stars
- 1
- 7-day trend score
- 8
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/moonshineaitech/lunastack/perf-budget)