AgentEval maf-updates.instructions.md

A set of instructions for updating Microsoft Agent Framework, a .NET framework for building AI agents, inside AgentEval. It guides package updates, breaking-change fixes, and tests.

In plain words
What is it for?
It helps check an existing upgrade plan, locate pinned package versions, adapt AgentEval to API changes, and verify the result with builds and tests.
Why use it?
It provides a repeatable way to find and handle code that stops working after a framework version changes.

Instructions file for GitHub Copilot

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 instructions/agentevalhq/agenteval/maf-updates
Clone the repo
git clone --depth 1 https://github.com/AgentEvalHQ/AgentEval

Made for: GitHub Copilot.

Per session 3,171 This file is loaded in full into every session.
When invoked 3,171 The same file — it is already loaded in full.
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 $0.03171 $0.03171
Opus 5 $0.01586 $0.01586
Sonnet 5 $0.00634 $0.00634
Haiku 4.5 $0.00317 $0.00317

Measured yesterday against content hash bc76726b8a6e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

AgentEval maf-updates.instructions.md 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 yesterday.

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.

.github/instructions/maf-updates.instructions.md · 248 lines

How it starts

The opening of the file, as written. The whole thing — 248 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MAF Update & Breaking Change Instructions

These instructions apply when updating the Microsoft Agent Framework (MAF) NuGet dependency in AgentEval.

Two-Phase Upgrade Workflow

MAF upgrades follow two phases:

Phase 1 (OPTIONAL):  maf-upgrade-preparation.instructions.md
                     Diff MAF source → produce plan document
                     Output: MAF/MAF-Upgrade-Plan.md

Phase 2 (THIS FILE): Update NuGet → fix breaks → run tests
                     Input: plan document (if available) OR compile errors

Phase 1 is recommended for major MAF version changes. For minor/patch updates or Dependabot PRs, start directly at Step 0 below.

Step 0: Check for an Existing Upgrade Plan

Before starting, check if MAF/MAF-Upgrade-Plan.md exists.

  • If it exists: Read it first. It contains pre-analyzed breaking changes, behavioral changes, and concrete fix code snippets produced by Phase 1. Use it as a guide — you already know what will break and how to fix it.
  • If it doesn't exist: Proceed normally with build-error-driven detection (Step 1).

Where MAF Versions Are Pinned

All MAF package versions are pinned in a single file:

Directory.Packages.props

Look for these three entries:

<PackageVersion Include="Microsoft.Agents.AI" Version="..." />
<PackageVersion Include="Microsoft.Agents.AI.OpenAI" Version="..." />
<PackageVersion Include="Microsoft.Agents.AI.Workflows" Version="..." />

All three MUST be updated to the same version simultaneously.

Also update the Microsoft.Extensions.AI packages if the MAF release requires a newer version:

<PackageVersion Include="Microsoft.Extensions.AI" Version="..." />
<PackageVersion Include="Microsoft.Extensions.AI.OpenAI" Version="..." />
<PackageVersion Include="Microsoft.Extensions.AI.Evaluation.Quality" Version="..." />

Current versions — always check Directory.Packages.props for the actual pinned versions.

Which projects reference which MAF packages:

  • src/AgentEval.MAF/AgentEval.MAF.csprojMicrosoft.Agents.AI, Microsoft.Agents.AI.Workflows (compile-time dependency)
  • src/AgentEval/AgentEval.csproj (umbrella) → Microsoft.Agents.AI, Microsoft.Agents.AI.Workflows (re-declared for NuGet consumers because PrivateAssets="all" suppresses transitive propagation from sub-projects)
  • samples/AgentEval.Samples/AgentEval.Samples.csprojMicrosoft.Agents.AI.OpenAI, Microsoft.Agents.AI.Workflows

Read the full file on GitHub · 248 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. yesterday First seen · 248 lines · 3,171 tokens per session scan A bc76726b8a6e

Subscribe to this mod's changes

AgentEval maf-updates.instructions.md is an instructions file published in the GitHub repository AgentEvalHQ/AgentEval (138 stars, last pushed yesterday), licensed MIT. It adds 3,171 tokens to every session, about $0.0159 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 2026-08-30.