ThirdKeyAIThirdKeyAIOfficialยท1 Agent Skills Included

Symbiont

Policy-governed runtime for secure, auditable autonomous agents

Runs autonomous agents under enforced security policies, sandbox isolation, and cryptographic identity checks. Eliminates blind trust in tool calls with Cedar policy gates, declarative tool contracts, and tamper-evident audit logs. Scaffolds, tests, and deploys governed agents via one CLI, with Docker, gVisor, and Firecracker isolation tiers.
npx skills add ThirdKeyAI/Symbiont --all -g -y
Available:

Gives your AI agent the project structure, build commands, DSL syntax, sandbox tiers, and security rules it must follow when writing or running Symbiont agents.

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Symbiont?โ–ผ

Run `npx skills add ThirdKeyAI/Symbiont --all -g -y` in your terminal to install all skills in this suite globally.

What does Symbiont do?โ–ผ

Symbiont is a Rust runtime that executes AI agents under explicit policy, identity, sandboxing, and audit controls. It blocks unauthorized tool calls with Cedar policies instead of trusting prompts.

Can I run Symbiont without a cloud API key?โ–ผ

Yes. Point OPENAI_BASE_URL at any local OpenAI-compatible server like Ollama, or use the offline policy evaluation and tool dry-run commands with no key at all.

How does Symbiont sandbox untrusted code?โ–ผ

You pick an isolation tier per agent: Docker by default, gVisor for untrusted code, or Firecracker microVMs for regulated data. All tiers ship in the open-source release.

Does Symbiont work with Claude Code?โ–ผ

Yes. Symbiont can spawn Claude Code as a governed subprocess with a policy-gated launch, tool allowlists, budget limits, and a live audit journal of every tool call.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’