prompt
FreeUnified virtual filesystem design for AI agents
FreeFree tier
About prompt
A design blueprint prompt for AI agents tasked with creating a unified virtual filesystem layer. It instructs the agent to architect a system that maps heterogeneous backends (S3, Google Drive, Slack, Gmail, Redis, GitHub, databases, APIs) into a single mount tree accessible via Unix-like commands (cat, cp, grep, find, ls, wc, jq). The prompt covers mount topology, resource adapters, tool surfaces, caching, portability, and integration with agent frameworks like OpenAI Agents SDK, Vercel AI SDK, LangChain, and Pydantic AI.
Key Features
Unifies multiple backends (S3, Google Drive, Slack, Gmail, Redis, GitHub, databases) into a single mount tree
Uses standard Unix commands (cat, cp, grep, find, ls, wc, jq) for agent interaction
Two-layer cache with TTL and invalidation per backend
Pluggable cache backends (RAM, Redis, disk)
Workspace snapshots for portability across machines
MCP bridge exposes mounts as MCP resources/tools
System-prompt hints to teach agents the mount layout
Integrates with OpenAI Agents SDK, Vercel AI SDK, LangChain, Pydantic AI
Pros & Cons
Pros
- Collapses N APIs into one familiar Unix-like abstraction
- Leverages agents' existing fluency in bash vocabulary
- Two-layer caching improves performance across backends
- Portable workspace snapshots allow cloning across machines
- Designed to integrate with major agent frameworks and MCP
Cons
- Requires significant upfront design and implementation effort
- Not a ready-to-use tool; it is a design prompt for agents
- May introduce overhead for simple tasks that could use direct API calls
Best For
Designing a unified filesystem abstraction for AI agents that interact with multiple APIsReducing complexity of integrating N SDKs and MCPs into one familiar abstractionBuilding agent workflows that mix data from cloud storage, messaging, and databasesEnabling agents to perform cross-backend operations like copying from S3 to a staging directory
FAQ
What is the purpose of this prompt?
It instructs an AI agent to act as a senior virtual filesystem architect, designing a unified layer that maps multiple backends into a single mount tree accessible via Unix commands.
Which backends are mentioned?
S3, Google Drive, Slack, Gmail, Redis, GitHub, databases, and APIs.
Does this prompt provide actual code?
No, it is a design specification. It outlines responsibilities and design principles for the agent to produce a system design.
What agent frameworks are targeted?
OpenAI Agents SDK, Vercel AI SDK, LangChain, and Pydantic AI.