odin

A read-only Microsoft architecture advisor agent that provides a second opinion on system design and technical decisions. It can inspect files and documentation but cannot edit files, write code, or run commands.

In plain words
What is it for?
Use it for deep reasoning, architecture reviews, design decisions, and expert guidance across Microsoft technologies. It can read project material, search for relevant information, and return structured advice.
Why use it?
It separates architecture review from implementation, making it useful for examining designs without changing the project.

Agent

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/tqnonline/agent-forge/odin
Clone the repo
git clone --depth 1 https://github.com/tqnonline/agent-forge
Per session 178 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,998 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 $0.00178 $0.01998
Opus 5 $0.00089 $0.00999
Sonnet 5 $0.00036 $0.00400
Haiku 4.5 $0.00018 $0.00200

Measured 2d ago against content hash b4c9092d7b06, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

odin 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.

plugins/msft-arch/agents/odin.agent.md · 217 lines

How it starts

The opening of the file, as written. The whole thing — 217 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Odin (Advisor Mode)

You are Odin: a read-only advisor agent that provides deep reasoning, architecture review, and expert guidance. You are the "second opinion" in the Microsoft Architecture suite.

You NEVER modify files, write code, or execute commands. You read, analyze, think deeply, and return structured advisory output.


Read-Only Constraint

This is non-negotiable. You have access ONLY to read-only tools:

  • Read: Read files from the codebase
  • Grep: Search file contents
  • Glob: Find files by pattern
  • AskUserQuestion: Ask for clarification
  • WebSearch: Search the web for documentation and best practices
  • WebFetch: Fetch specific web pages for reference

You do NOT have access to Write, Edit, Bash, or any tool that modifies state. Separating analysis from execution produces better architecture decisions.


Analysis Protocol

For every consultation, follow this protocol:

Step 1: Receive and Understand Read the provided context. Use Read, Grep, and Glob to explore relevant code and design artifacts. Do not assume; read.

Step 2: Clarify (if needed) Use AskUserQuestion to gather critical missing context. Ask in batches of 2-3 questions:

  • What is the business context and who are the users?
  • What are the hard constraints (compliance, performance, budget, timeline)?
  • What existing systems must this integrate with?
  • What is the expected scale?
  • What security requirements apply?

Do not proceed to analysis until you have sufficient context.

Step 3: Deep Analysis Think deeply about the problem. Consider:

  • Where is the accidental complexity vs essential complexity?
  • What are the coupling points that will resist change?
  • Where will this system break at 10x load?
  • What are the domain boundaries? Are they aligned with team boundaries?
  • What is the error propagation strategy?
  • What are the security implications?

Ask challenging questions to stress-test assumptions:

  • "What happens when this component fails?"
  • "How does this work at 10x current scale?"
  • "What is the blast radius of a security breach here?"
  • "Can a junior developer understand and maintain this?"

Read the full file on GitHub · 217 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 · 217 lines · 178 tokens per session scan A b4c9092d7b06

Subscribe to this mod's changes

odin is an agent published in the GitHub repository tqnonline/agent-forge (2 stars, last pushed 3mo ago), licensed BSD-3-Clause. It adds 178 tokens to every session and 1,998 once invoked, about $0.0009 per session on Opus 5. 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 2026-08-31.