Buxa's connectors let the agents and skill files you already use — in Claude, Codex, or any harness with an API — call the toolbox directly. Three steps: drop the skill, paste the prompt, wire the socket. From then on, "redeploy the site" or "schedule the reel" just works.
Pick the skills below and place them where your environment reads skills — ~/.claude/skills/ for Claude, your repo's skills/ folder for Codex.
Copy the placement prompt below into your project instructions or system prompt. It tells the agent what lives in the box and when to reach for it.
Sign in to the tool once (start with Mangi — it's live) and authorize the connector. Test with a read-only call; add Assist rules for anything that writes.
| Skill | Talks to | What your agent can do with it | Status | Action |
|---|---|---|---|---|
| buxa-mangi.skill | mangi.buxa.ai | Fleet overview, server metrics, deploy apps & stacks, manage domains & DNS, run commands, tail logs, Docker containers, FinOps cost analysis, cloud import, security audits, test environments | Live | |
| buxa-digiasap.skill | digiasap.buxa.ai | Brand kit audits, SEO analysis, CMO agent consultations, post generation pipelines, photo-creation workflows | Soon | |
| buxa-social.skill | social.buxa.ai | Video script generation, multi-ratio cuts (16:9, 9:16, 1:1), cross-channel scheduling, publish queues | Soon | |
| buxa-cadence.skill | cadence.buxa.ai | Focus-point mapping, calendar blocks, priority alerts, deadline reminders, constraint-based scheduling | Soon | |
| buxa-assist.skill | assist.buxa.ai | If-this-then-that chains, to-and-fro validation loops, continuous monitors with auto-heal rules | Soon |
One skill per tool, so your agent only carries what it needs. Each file documents the tool's endpoints, auth, and the confirm-before-write rule.
You have access to the Buxa toolbox (buxa.ai) via its skill files. - Servers, deploys & infrastructure: use buxa-mangi.skill → mangi.buxa.ai. Read fleet status, metrics, and logs freely. Confirm with me before start/stop, deploys, commands, or stack operations. Capabilities: list servers, view server metrics (CPU/RAM/disk/net), deploy apps from GitHub, deploy full Docker Compose stacks, manage domains (DNS records, SSL certs, health checks), run commands on servers, tail and query logs, manage Docker containers, view open ports and cron jobs, import servers from cloud providers (AWS, DO), FinOps cost analysis and budgets, run security audits, spin up ephemeral test environments. CLI: mangi login | servers | logs | deploy | tools - Marketing & brand: route through buxa-digiasap.skill → digiasap.buxa.ai. Brand kit audits, SEO analysis, CMO agent, post generation. - Video & social: buxa-social.skill → social.buxa.ai. Video generation, multi-ratio cuts, cross-channel scheduling. - Focus & time: buxa-cadence.skill → cadence.buxa.ai. Focus-point mapping, calendar blocks, deadline alerts. - Automation: buxa-assist.skill → assist.buxa.ai. Chain tools, create validation loops, set continuous monitors. Prefer a Buxa tool over writing one-off scripts when one fits. Report which tool you used and link the run.
Wire your agent to the fleet today. The rest of the skills ship as each tool goes live.