Skip to content
UsecaseBoard
← All use cases
githubagent-sandbox

Perplexity SPACE: open-source Firecracker sandboxes for agent code

At a glance

Perplexity open-sourced SPACE (Sandboxed Platform for Agentic Code Execution)—the Firecracker microVM sandbox behind Computer—under Apache 2.0 on Oct 15, 2026: HTTP/SSE APIs to create Linux environments, run commands, and manage lifecycle; partners CoreWeave, Akamai, and Crusoe plan to offer it; Semafor reports 0/118 breakout attempts vs other sandboxes in Perplexity’s tests. This workflow is listed under agent-sandbox, open-source, security, Firecracker, governance.

Agents / tools
Perplexity SPACE, Perplexity Computer, Firecracker

Building a tool for these workflows? Sponsor / inquire.

Save ideas in this browser. Upvotes count across the board.

Prompt excerpt

Deploy SPACE in a local Kind cluster per the docs, create one sandbox from a template, run a command, and confirm the agent cannot reach host credentials.
Explore all →
githubobservability

Tracekit 1.2.3: tamper-evident flight recorder for AI agents

MIT Python package signs every policy-gated tool call into a hash-chained .tkb file verifiable offline (Python+TS); works across Claude Code, OpenAI Agents, LangChain/LangGraph, MCP and more—agents can edit their own logs but not hide edits when the signer is out of reach.

TracekitClaude CodeOpenAI Agents+3
Source ↗
docsGrok-Bot

Agent Vault: Grok Bot and Meta Muse shop/login without seeing your card

Self Labs’ Agent Vault gives agents mock tokens for checkout and sign-in; real card numbers and passwords never reach the model. Caps, merchant allowlists, and per-purchase approval. Connect Grok Bot via MCP or Meta Muse via custom connector in ~2 minutes (also Claude Code, Hermes, Codex).

Agent VaultGrok BotMeta Muse+4
Source ↗
docsagentic-SDLC

Anthropic AI-native SDLC playbook: intent.md → hooks → closed loop

Anthropic’s Applied AI playbook reframes Plan/Design/Build/Test/Deploy/Maintain as a loop of committed artifacts (intent.md → spec.md → plan.md → PR → production) with Claude Code plan mode, skills as institutional knowledge, hooks as deterministic gates, continuous evals in CI, and production control bands that write the next intent.md. Humans stay at judgment gates; agents do the rest. Companion pieces cover securing an AI-native SDLC and Claude Academy course modules.

Claude CodeClaude Agent SDKhooks+2
Source ↗
githubagent-sandbox

Managoat: HTTP API that runs Claude Code, Codex, or Gemini CLI on parked sandboxes

Managoat turns coding agents into a conversational computer API: POST a prompt, a sandbox wakes with repos/packages/credentials, Claude Code or Codex or Gemini CLI runs, then the machine parks (disk kept, no awake compute) until the next turn. Agent/Environment/Vault templates; secrets merge with Vault winning; bound secrets never enter the sandbox (egress broker). REST API, CLI, TypeScript SDK, demos (Conversations + Workbench), self-host docs. GitHub managoat/managoat actively shipping Oct 17.

ManagoatClaude CodeCodex+1
Source ↗