bama
BAMA: Build Any Mac App. Agent skills to research, clone, design, build, optimize, and test native macOS apps with Swift and SwiftUI.
翻譯準備中
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證Claude Code plugin marketplace and index of arm-pz agent skills: editorial, teaching, recruiting, SEO planning, release preflight, sub-agent tournaments and repo discovery. Install with npx skills add arm-pz/caffeinity-skills.
翻譯準備中
取自儲存庫 README
正在載入完整 README。
檢視完整 README ↗Agent skills for Claude Code, Codex, Cursor and any agent that reads a SKILL.md, by Abdul Rahman P Z:
44 skills for writing, SEO, release checks, skill authoring, careers, decisions, learning and
outbound, plus eleven workflow agents that hand work from one skill to the next. Browse them at
skills.caffeinity.com.
Claude Code — the caffeinity plugin bundles every skill and the eleven agents:
/plugin marketplace add arm-pz/caffeinity-skills
/plugin install caffeinity@caffeinity-skills
Codex — the same plugin, from this repo's Codex marketplace (.agents/plugins/marketplace.json):
codex plugin marketplace add arm-pz/caffeinity-skills
then install caffeinity from the Plugins directory.
Any other agent (Cursor, Qoder, Gemini CLI, …) — one skill at a time with the skills installer:
npx skills add arm-pz/<skill>
Install either the bundle or individual skill plugins, not both, or the same skill loads twice.
Each agent runs one workflow and preloads the skills it needs. In Claude Code, ask for it by name
(@caffeinity:seo-lead) or let Claude delegate when a request fits its description.
| Agent | Workflow | Skills, in order |
|---|---|---|
skill-builder | Brew 1 · Pour-over: Build a skill and make it installable | draft-skill or prompts-to-skill → tune-skill → prompts-to-skill-release → skill-to-mod |
discovery-lead | Brew 2 · Bloom: Get a repo, skill or plugin found | findable → publish-pipeline → upstream-check |
release-gatekeeper | Brew 3 · Espresso: Decide whether a site is safe to launch | web-preflight → pre-ship-security-red-team → pre-ship-performance-audit → release-gate-chair |
seo-lead | Brew 4 · Cold brew: Grow organic search traffic | seo-crawl → keyword-guard → wordsmith → pre-ship-performance-audit / platform-content |
job-search-coach | Brew 5 · French press: Land a job | recruiter-lens → applicant-kit → application-pilot → career-track |
decision-analyst | Brew 6 · Cupping: Make a decision | research-brief → data-readout → crash-test or do-better → strategy-memo |
learning-coach | Brew 7 · Slow drip: Learn a skill | topic-primer → study-system → course-builder |
outbound-lead | Brew 8 · Nitro: Run cold outbound | cold-email-list-gate → cold-email-ops → sales-copy |
solo-launcher | Brew 9 · Moka pot: Launch a one-person product | solo-income-blueprint → site-prompt → production-code → paper-trail → release-gate-chair → platform-content |
skill-maintainer | Brew 10 · Descale: Maintain the skill family | tune-skill → prompts-to-skill-release → upstream-check |
estate-lead | Brew 11 · Estate roast: Stand up or audit a skill estate | skill-estate |
Codex has no subagents, so there the caffeinity-routing skill does this job: describe the outcome
and it names the workflow and the next skill.
The bundle is generated, not edited: scripts/build_plugin.py renders the agents from
plugins/caffeinity/agents.json and caffeinity-routing's workflows.md, and copies each skill
from its own repo (plugins/caffeinity/SOURCES.md records the commits). The build plugin
workflow reruns it hourly and bumps the version only when something changed.
evals/triggers.json holds one or more real user messages per skill, with the skill that must load
and the ones that must not. scripts/build_evals.py turns it into a claude plugin eval suite under
plugins/caffeinity/evals/ (CI checks it is current and that every bundled skill has a case), and
bash scripts/run_evals.sh
runs it on your own Claude login: about $0.10 a case, $5 for the suite, never in CI. Each case checks
which skill the Skill tool actually loaded. evals/last-run.json records the result per case; a rerun
prints NEWLY FAILING and NOW PASSING so a description edit that breaks triggering is caught by name.
Single runs vary, so treat a case as a real miss when it fails twice.
/plugin marketplace add arm-pz/caffeinity-skills
/plugin install wordsmith@caffeinity-skills
Replace wordsmith with any of: caffeinity-routing, recruiter-lens, applicant-kit, topic-primer, application-pilot, do-better, keyword-guard, seo-crawl, web-preflight, findable, crash-test, pre-ship-performance-audit, pre-ship-security-red-team, release-gate, publish-release-gate, upstream-check, tune-skill, draft-skill, prompts-to-skill, prompts-to-skill-release, publish-pipeline, cold-email-ops, cold-email-list-gate.
Each of those plugins lives in its own repository (see .claude-plugin/marketplace.json for sources).