Browse categories
Explore
Fiverr Pro
English
$
USD
I will setup, build, deply hermes agent openclaw mcp claude code obsidian ai docker
Adakole Emmanue
About this gig
Your agent shouldn't be a security hole, a runaway token bill, or a lost weekend.
Most self hosted agents fail on three things: a broken Docker install, an unpatched CVE, and no memory between sessions. I fix all three.
WHAT YOU GET
- OpenClaw or Hermes Agent, self hosted on Docker or VPS
- OpenClaw security audit: CVE-2026-25253 unauthenticated RCE (CVSS 8.8) patched
- ClawHub plugins + OpenClaw plugins installed and version pinned
- Hermes agent OpenClaw migration, or both running side by side
- Hermes agent memory, SOUL.md identity, Hermes cron scheduled tasks
- MCP wired: Obsidian MCP, Claude Code, Claude Cowork, Claude for Excel
- Model routing: Opus 5 API, Kimi K3, Moonshot AI open weight, local Ollama
- Obsidian second brain: Obsidian AI plugins, Copilot for Obsidian, Obsidian AI workflow
- Hermes Atropos fine tuning and Hermes agent DGX Spark builds on request Free decision doc: OpenClaw vs Hermes, OpenClaw alternatives, Jasper AI alternatives pricing, licensing, Frontier-Bench and GDPval-AA notes.
- Revisions until it runs. Message your stack first.
You'll receive a secure, documented and expandable AI workspace designed for development, automation and long-term reliability.
Get to know Adakole Emmanue
Adakole Emmanue
Minecraft Hytale and Modpack Fixes That Work
- FromNigeria
- Member sinceApr 2026
- Avg. response time1 hour
- Last delivery1 month
Languages
English, Spanish, German, French, Dutch
I help gamers turn broken, laggy, or unfinished game setups into smooth, stable experiences they can actually enjoy. I can create, fix, update, and improve Minecraft modpacks and servers, including Cobblemon, Forge, Fabric, and Hytale projects. Whether your game crashes, runs slowly, or simply isn’t working as expected, I’ll find the problem, fix it, test everything, and make sure you’re ready to play. Clear communication, reliable work, and support from start to finish.
My Portfolio
FAQ
Should I choose OpenClaw or Hermes Agent?
Depends on your bottleneck. OpenClaw wins on channel breadth and the ClawHub plugins ecosystem. Hermes agent memory and self-improving skills win when the same workflow repeats weekly. I send an OpenClaw vs Hermes comparison before you order, free, no obligation. Many clients end up running both.
Is OpenClaw safe to self host after the CVEs?
Yes, when configured properly. CVE-2026-25253 was an unauthenticated RCE rated CVSS 8.8, and thousands of exposed instances were hit. Every OpenClaw self hosted build I ship binds to localhost, enforces token auth, locks file permissions, and runs behind a reverse proxy.
Can you run this in Docker or on my own VPS?
Yes. OpenClaw Docker and Hermes both deploy to any Linux VPS, Mac Mini, Raspberry Pi, or DGX box. I harden the container, set resource limits, configure auto-restart and backups, and hand you the compose file so you can rebuild it yourself anytime. No lock-in to me.
Do you handle OpenClaw to Hermes agent migration?
Yes, this is my most requested job. I port your skills, credentials, channel configs and history, rebuild them as Hermes skills with persistent memory, then run both in parallel until you confirm parity. Zero downtime. Rollback plan included. Hermes agent OpenClaw migration typically takes 2–4 days
What are ClawHub plugins and do I need them?
ClawHub is OpenClaw's plugin marketplace that replaced the older skills system. Most tutorials online still reference the old format, which is why installs break. I install OpenClaw plugins from ClawHub with pinned versions and dependency checks, so an upstream update never silently kills your agent
Can you set up MCP servers?
Yes. MCP is the connector layer that lets your agent reach real tools. I wire Obsidian MCP, Google Drive, Slack, GitHub, databases and custom endpoints, plus Claude Code, Claude Cowork and Claude for Excel where relevant. Every MCP server is tested with live calls before handover, not just configure
Which AI models can it run?
Any. Opus 5 API for balanced cost and reasoning, Kimi K3 or other Moonshot AI open weight models for cheap high-volume work, or fully local via Ollama for zero data egress. I set up routing rules so cheap models handle routine tasks and premium models only fire when complexity demands it.
How do you cut my token costs?
Model routing plus effort tuning. Opus 5 exposes an effort dial, and Kimi K3 and other open weight models cost a fraction for bulk work. I benchmark your actual tasks, then route accordingly. Clients typically see 40–70% lower spend versus running one premium model for everything.
Will the agent remember things between sessions?
Yes, that's the core of Hermes agent memory. I configure the layered memory stack, write your SOUL.md so the agent holds a consistent identity and tone, and enable procedural memory so successful workflows become reusable skills.
Can it run tasks automatically without me prompting it?
Yes. Hermes cron scheduled tasks and OpenClaw's heartbeat scheduler both handle this. Daily briefings, inbox triage, invoice chasing, deploy monitoring, report generation, each runs in a fresh session and delivers output to your chosen channel. I set the schedules and failure alerts.
