agent/agent-framework
1
0
Fork 0
mirror of https://github.com/microsoft/agent-framework.git synced 2026-06-16 21:04:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

Capture meeting notes

Browse Source
This commit is contained in:
Chris Rickman
2026-03-17 13:34:48 -07:00
Unverified
parent 5f7bca0963
commit f5d4da9f0e
1 changed files with 105 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/harness/pipeline.md
+105 -25
View File
@@ -1,3 +1,26 @@
# MAF Brainstorming Session
> March, 17th 2026
## Reference Documents
- [OpenClaw Agent Harness](https://microsoft-my.sharepoint.com/:w:/p/shahen/IQAU5F524RvtTpzjAZIjMf0BAQqMMPYCMRRPnX0eslRiPb0?e=MLF6CL)
- [Agent Platform Comparison](https://m365.cloud.microsoft/chat/pages/eyJ1IjoiaHR0cHM6Ly9taWNyb3NvZnQuc2hhcmVwb2ludC5jb20vY29udGVudHN0b3JhZ2UveDhGTk8teHRza3VDUlgyX2ZNVEhMYmRXU2tHOE93Skt2VTBCX3pNTDFaVT9uYXY9Y3owbE1rWmpiMjUwWlc1MGMzUnZjbUZuWlNVeVJuZzRSazVQTFhoMGMydDFRMUpZTWw5bVRWUklUR0prVjFOclJ6aFBkMHBMZGxVd1FsOTZUVXd4V2xVbVpEMWlKVEl4TUhscVMzWlphRU5CVlVkV2IwUndhVlpzZUZKekxXTkNZMjltVm1jME1VSnlabVZpWWxKeVNXdFZlamd5YlRBelJtVnhXVlJaVmtwQlkzQlVkMnc0TUNabVBUQXhXVlJZVWxWVVNqSk1WbFV6UlZsWE4xTmFSek5JVDBGUFRWSkNXRWhLUzFFbVl6MGxNa1ltWVQxTWIyOXdRWEJ3Sm5BOUpUUXdabXgxYVdSNEpUSkdiRzl2Y0Mxd1lXZGxMV052Ym5SaGFXNWxjaVo0UFNVM1FpVXlNbmNsTWpJbE0wRWxNakpVTUZKVVZVaDRkR0ZYVG5saU0wNTJXbTVSZFdNeWFHaGpiVlozWWpKc2RXUkROV3BpTWpFNFdXbEZkMlZYY0V4a2JHeHZVVEJHVmxJeFduWlNTRUp3Vm0xNE5GVnVUWFJaTUVwcVlqSmFWMXA2VVhoUmJrcHRXbGRLYVZWdVNrcGhNVlkyVDBSS2RFMUVUa2RhV0VaYVZrWnNWMU5yUm1walJsSXpZa1JuZDJaRVFYaFhWbEpaVld4V1ZWUldUa0pPTUdSWlZGVk9SbFpVVWs5U2EydDVVMVJqZVZGc2FGWldSa3BDVFRGSkpUTkVKVEl5SlRKREpUSXlhU1V5TWlVelFTVXlNbU5oTkRSbFlUZGtMVEF5TWpjdE5ERm1OQzFpWVRsaUxUUm1OV1kyTlRWaVpEZGpPU1V5TWlVM1JBPT0ifQ?auth=2&ct=1773767195553&or=Teams-HL&LOF=1)
- [LangChain Deep Agents](https://docs.langchain.com/oss/python/deepagents/overview)
- [CodeAct](https://arxiv.org/abs/2402.01030)
- [AI Accelerator - Foundry](https://microsoft.sharepoint.com/:p:/t/CoreAIStudioOutboundProduct/IQAFfbFIu5E7RLMcFmX_7C6EAU2xt1TxnGbR36EN6JTwX0Y?e=9UsKZ6)
- [Foundry Developer Portal (Hosted Agents)](https://hosted-agents-builder.lemonriver-6a2ef1ee.westus2.azurecontainerapps.io/getting-started)
## Next Steps
Demo for MVP session on campus next week.
1. Code Act: How to with MAF
1. Harness Preview: Single agent with compaction, tools (including shell or file-system), and simple orchestration loop.
Deploy agent with harness as _Foundry Hosted Agent_.
## What Is the Agent Harness?
The runtime control plane that enables reliable, long-running agent execution. Not a specific agent. Not a specific set of tools. It's the **infrastructure layer** that any agent can run within.
@@ -11,29 +34,51 @@ What it provides:
- **Context management** — compaction, eviction, externalization
- **Observability** — traces, transcripts, replay
### Where We Stand
MAF compared against DeepAgents, Amplifier, Opencode, Copilot CLI, OpenAI Codex, and Claude Code — every one of them has these capabilities. MAF has some partially, most not at all.
| Priority | Capability | Layer | MAF Today |
|---|---|---|---|
| **P0** | Agent/User Orchestration | Harness | Workflows |
| **P0** | Filesystem tools | Environment | Local access |
| **P0** | Shell execution | Environment | Local access |
| **P0** | Context compaction | Harness | Pipeline + initial strategies |
| **P0** | Todo / planning tool | Harness | Nothing |
| **P0** | Sub-agent delegation | Harness | Partial — orchestration exists, state isolation incomplete |
| **P0** | Memory | Environment | Partial — Python core only |
| **P1** | Skills / prompt presets | Persona | Nothing |
| **P1** | Model routing | Harness | Nothing |
|---|:---|:---|:---|
| **P0** | Agent/User Orchestration | Harness | 🟡 Workflows or part of `AIAgent`? |
| **P0** | Session Persistence | Harness | 🟢 `AgentSession`/`AIContextProvider` |
| **P0** | Context Compaction | Harness | 🟢 In preview |
| **P0** | Memory | Environment | 🟡 Partial — Python core only |
| **P0** | Permissions / Scoping | Harness | 🔴 Nothing |
| **P0** | Tool: Composite Tool Calling | Harness | 🔴 Needs definition |
| **P0** | Tool: Filesystem | Environment | 🟢 Local access |
| **P0** | Tool: Shell Execution | Environment | 🟢 Local access |
| **P0** | Tool: Todo / Planning | Harness | 🔴 Nothing |
| **P0** | Sub-Agent Delegation | Harness | 🟡 Partial — orchestration exists, state isolation incomplete |
| **P1** | Skills / Prompt Presets | Persona | 🔴 Nothing |
| **P1** | Model Routing | Harness | 🔴 Nothing |
| **P1** | Agent Budgets | Persona | 🔴 Nothing |
| **P1** | Prompt Caching | Persona | 🔴 Nothing |
### Open Issues
- How does this look in DevUI?
---
## Features
### Agent/User Orchestration
How does the harness manage structured, multi-turn data collection from the user? What's the interaction model between the outer loop and user-facing slot-filling prompts? How does it compose with compaction and task management? How does the agent re-ask or repair slot values after partial completion?
How does the harness manage structured, multi-turn data collection from the user?
What's the interaction model between the outer loop and user-facing slot-filling prompts?
How does it compose with compaction and task management? How does the agent re-ask or repair slot values after partial completion?
### Session Persistence
`AgentSession` is directly serializable and there is also a `ChatHistoryProvider` option for more complex storage needs.
Open questions:
- How to handle schema evolution?
- How to version `AgentSession`?
- How to support partial loading for long histories?
- Does a session include more than conversation context (i.e. messages)?
### Compaction Strategy
@@ -41,15 +86,41 @@ Two API tiers — **simple for most developers, advanced for full control**.
**Simple (menu-driven)** — developer picks from preset enums:
```csharp
harnessBuilder.AddCompaction(Approach.Balanced, Size.Compact, summarizingChatClient);
builder.AddCompaction(Approach.Balanced, Size.Compact, summarizingChatClient);
```
**Advanced (pipeline)** — ordered stages, least to most aggressive:
```csharp
builder.AddCompaction(
new PipelineCompactionStrategy(
// 1. Gentle: collapse old tool-call groups into short summaries
new ToolResultCompactionStrategy(CompactionTriggers.MessagesExceed(7)),
// 2. Moderate: use an LLM to summarize older conversation spans into a concise message
new SummarizationCompactionStrategy(summarizerChatClient, CompactionTriggers.TokensExceed(0x6000)),
// 3. Aggressive: keep only the last N user turns and their responses
new SlidingWindowCompactionStrategy(CompactionTriggers.TurnsExceed(32)),
// 4. Emergency: drop oldest groups until under the token budget
new TruncationCompactionStrategy(CompactionTriggers.TokensExceed(0x8000))));
```
1. **Gentle:** Collapse old tool-call groups into short summaries (`ToolResultCompactionStrategy`)
2. **Moderate:** LLM-based summarization of older conversation spans (`SummarizationCompactionStrategy`)
3. **Aggressive:** Sliding window — keep only last N user turns (`SlidingWindowCompactionStrategy`)
Open questions: What are the right defaults? How do triggers compose when multiple strategies are pipelined?
### Memories
Loads from backend storage, injects into system prompt automatically.
### Permissions / Scoping
Agent Scopes define the explicit boundaries within which an agent is allowed to operate, constraining where it can act
(for example, a single folder or service) and what level of access it has (such as read‑only, write, or execute).
By enforcing scoped resources and permissions, this feature ensures the agent’s actions remain intentionally limited,
predictable, and aligned with least‑privilege principles—preventing overreach even when the agent could otherwise reason
about broader options.
### Tools: File System / Shell
@@ -61,7 +132,7 @@ Both should follow the same pattern: **interface-based, with pluggable backends*
- **Computer Use (TODO)** — same interface pattern for screen/mouse/keyboard interaction. Future work. Security and governance implications are significant.
### Task Management
### Tools: Task Management
`TodoTool` with `write_todos`. `TodoItem` has content + status (pending / in progress / completed). `TodoMiddleware` injects current todos into the system prompt — this is what gives the agent the ability to self-plan and track its own progress.
@@ -70,11 +141,6 @@ Open debate: P0 or P1? "An agent _can_ work with only filesystem + shell. But fo
### Data Driven: Input Schema / Structured Data Output
How does the harness support defining what data the agent needs (input schema) and what the agent produces (structured output)? Is this related to or distinct from slot filling? How does the developer define and validate schemas?
### Memories
Loads from backend storage, injects into system prompt automatically. Open questions: How does memory interact with compaction — should compacted summaries become long-term memories? Which backends out of the box? What's the cross-session story?
### Sub-Agent Delegation
MAF already has orchestration (Sequential, Concurrent, Group Chat, Magentic, Handoff, Human-in-the-loop). The gap is **state isolation**: sub-agents need their own message history and todo state, isolated from the parent, returning results as tool responses. "A lot of task planners and coding agents use sub-agents. This feels P0."
@@ -87,14 +153,30 @@ MAF already has orchestration (Sequential, Concurrent, Group Chat, Magentic, Han
`ModelRouterMiddleware` with strategies: cost-aware (minimize cost for task requirements) and heuristic (rule-based, e.g., stronger model for code tasks).
---
### Agent Budgets
Agent Budgets define a hard execution limit—such as a maximum number of tokens, turns, or tool calls—within which an agent must
decide whether to act and how far to pursue a task, directly shaping planning, delegation, and early stopping behavior.
Unlike a context‑management budget, which governs what information is retained or loaded, an agent budget constrains execution itself,
informing decisions like skipping steps, reducing depth, or terminating when the remaining budget cannot justify further action.
### Prompt Caching (P1)
Prompt caching is important for coding agents.
When an agent is iterating on a coding problem, the same or similar prompts are often repeated.
We need a great caching story to speed up iteration and reduce costs.
## Shape
Everything hangs off a **fluent builder pattern**:
Everything hangs off a **fluent builder pattern**.
Ideally this builder is identical with the agent-builder pattern, so developers
can seamlessly transition from "building an agent" to "building a harness for that agent" without learning a new API.
Example:
```
harnessBuilder
builder
.AddCompaction(...)
.AddTool(filesystemTool)
.AddTool(shellTool)
@@ -110,7 +192,6 @@ harnessBuilder
- _Production (Foundry Hosted Agents):_ Managed containers, autoscaling, identity, observability, Teams / M365 Copilot / Web
- The interface-based tool abstractions (`LocalFilesystem` ↔ `HostedFilesystem`, `LocalShell` ↔ `HostedShell`) are what make the two-tier model work.
---
## Validation
@@ -119,7 +200,6 @@ harnessBuilder
- **Compaction testing** — verify different pipeline configurations produce correct and useful results
- **Test strategy** — unit tests, integration tests, model-in-the-loop evaluation, benchmarks
---
## Tutorials