skills.rest
ruvnetruvnetCommunity·1 Agent Skills Included

midstream

Real-time LLM stream analysis, gating, and pattern detection

Analyzes live LLM token streams as they arrive, enabling pattern matching, scoring, and intervention before responses complete. Eliminates the black-box wait for finished responses by gating, steering, and verifying output per chunk with nanosecond-precision scheduling. Ships six composable Rust libraries plus WASM bindings so the same analysis pipeline runs natively, in browsers, and at the edge.
133|39
↗
npx skills add ruvnet/midstream --all -g -y
Available:

Instructs the AI agent to batch all operations concurrently in single messages, spawn sub-agents via the Task tool, follow SPARC test-driven phases, and keep files organized in proper subdirectories.

All Skills in This Repository (1)

Pure Emerald Level Indicators
📦 In Repo
ruvnetruvnet

AIMDS

Detect AI manipulation threats using temporal analysis, vector search, and formal verification.

Community
Advanced

Frequently Asked Questions

FAQPage Schema
How to install MidStream?▼

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

What does MidStream do with LLM streams?▼

It treats the token stream as a live signal you can pattern-match, score, and gate while tokens are still arriving, using Rust libraries for sequence comparison, scheduling, attractor analysis, and meta-learning.

Can I use MidStream in the browser?▼

Yes. The Rust workspace compiles to WASM and is published as the @midstream/wasm npm package, so the same analysis code runs in browsers and edge workers.

Does MidStream work with Claude Code?▼

Yes. It includes a CLAUDE.md orchestration file and SKILL.md units that guide Claude Code agents through concurrent execution, SPARC workflows, and stream-chain pipelines.

Do I need the whole workspace or just one library?▼

Each of the six crates is published independently on crates.io, so you can add only what you need, such as temporal-compare for pattern matching or the scheduler for low-latency queuing.

Related Repositories in Software Engineering

View All in Software Engineering→
openclawopenclaw
388.4k

openclaw

Run a personal AI assistant across your devices and chat apps

Official📦 120 Skills
obraobra
280.0k

superpowers

Gives coding agents a disciplined workflow from idea to merged code

Community📦 15 Skills
facebookfacebook
248.4k

react

AI agent skills for building, testing, and porting React core

Official📦 14 Skills