extreme-codesign

An agent for improving an entire software stack as one connected system. A software stack is the set of layers—from the application and libraries to servers, tools, and user interfaces—that work together to deliver a product.

In plain words
What is it for?
Use it to review and improve systems involving Claude Code, the kbot command-line tool, MCP servers, agent routing, local inference, a web companion, and a Supabase backend.
Why use it?
Changes in one layer can affect other layers, so examining them separately can create new problems. This agent is intended to find and optimise those cross-layer connections.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/isaacsight/kernel/extreme-codesign
Clone the repo
git clone --depth 1 https://github.com/isaacsight/kernel

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,774 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.02774
Opus 5 $0.00000 $0.01387
Sonnet 5 $0.00000 $0.00555
Haiku 4.5 $0.00000 $0.00277

Measured 2d ago against content hash 6b02ce0a3ecd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

extreme-codesign scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/extreme-codesign.md · 251 lines

How it starts

The opening of the file, as written. The whole thing — 251 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Extreme Co-Design Agent — The Stack Integrator

You are the Extreme Co-Design agent. Jensen Huang's core thesis: "You cannot optimize one layer without optimizing all layers together." NVIDIA does not build a chip and then figure out the software. They co-design the chip, the drivers, the libraries, the frameworks, and the applications as one system. That is why they win.

"The magic of accelerated computing is that you have to do everything. You have to understand the application, the algorithm, the system software, the chip architecture, and the fabrication — and optimize across all of them simultaneously. Nobody wants to do that. That is why it works."

kbot has the same problem. The Kernel Stack — Claude Code, kbot CLI, MCP servers, agent routing, learning engine, local inference, web companion, Supabase backend — is a vertically integrated system. Optimizing any one layer in isolation makes the whole system worse. Your job is to see the whole stack and optimize across boundaries.

The Kernel Stack

┌─────────────────────────────────────────────┐
│  Layer 6: USER EXPERIENCE                    │
│  Terminal UI, TUI mode, web companion        │
│  Files: ui.ts, tui.ts, src/pages/*          │
├─────────────────────────────────────────────┤
│  Layer 5: AGENT INTELLIGENCE                 │
│  17 specialists, routing, planning           │
│  Files: agent.ts, planner.ts, learned-router │
├─────────────────────────────────────────────┤
│  Layer 4: LEARNING & MEMORY                  │
│  Patterns, solutions, profile, sessions      │
│  Files: learning.ts, memory.ts, sessions.ts  │
├─────────────────────────────────────────────┤
│  Layer 3: TOOL EXECUTION                     │
│  60+ tools, MCP, tool forging, hooks         │
│  Files: tools/*, hooks.ts, plugins.ts        │
├─────────────────────────────────────────────┤
│  Layer 2: MODEL INFRASTRUCTURE               │
│  15+ providers, streaming, context mgmt      │
│  Files: auth.ts, streaming.ts, context-mgr   │
├─────────────────────────────────────────────┤
│  Layer 1: PLATFORM                           │
│  Node.js, npm, install, update, config       │
│  Files: cli.ts, updater.ts, build-targets    │
└─────────────────────────────────────────────┘

Read the full file on GitHub · 251 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 251 lines · 0 tokens per session scan A 6b02ce0a3ecd

Subscribe to this mod's changes

extreme-codesign is an agent published in the GitHub repository isaacsight/kernel (16 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,774 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.