Servers overview

tuls packages six focused MCP servers into one binary:

Server	Purpose	Primary capability
filesystem	Read, inspect, search, edit, and move workspace files	filesystem.read, filesystem.write
fetch	Fetch bounded HTTP(S) content with explicit network policy	network.fetch
memory	Maintain a persistent JSONL knowledge graph	memory.read, memory.write
shell	Execute programs with direct argv semantics	process.execute
skills	Discover and activate workspace skills	skills.read
agents	Run local provider-backed subagents with child MCP tools	agents.run

Each server:

• speaks MCP 2026-07-28 over stdio;
• starts with a capability policy you control via --allow/--deny;
• bounds every network, file, and tool output it produces;
• rejects unknown JSON fields on public tool inputs.

filesystem→

Read, inspect, search, edit, and move workspace files inside explicitly allowed roots.

tuls filesystem /path/to/project --allow filesystem.read

filesystem.readfilesystem.write

fetch→

Bounded HTTP(S) fetches with robots.txt and public-network policy, no redirects by default.

tuls fetch --allow network.fetch

network.fetch

memory→

A persistent JSONL knowledge graph with entities, relations, and observations.

tuls memory --memory-file memory.jsonl --allow memory.read --allow memory.write

memory.readmemory.write

shell→

Execute one program directly with an argv array — no shell parsing, bounded output.

tuls shell /path/to/project --allow process.execute

process.execute

skills→

Discover and activate workspace skills with their instructions and resource manifests.

tuls skills /path/to/project --allow skills.read

skills.read

agents→

Run local provider-backed subagents with their own child MCP tool policies.

tuls agents /path/to/project --allow agents.run

agents.run