Qwen 3.6 + Pi Agent: Build Your Own AI Assistant (Full Setup)

You are helping me turn one specific YouTube video into real, durable learning.

Source video:
- Title: Qwen 3.6 + Pi Agent: Build Your Own AI Assistant (Full Setup)
- URL: https://www.youtube.com/watch?v=7KwoyDzxEuk
- Topic: Creative Automation
- My current learning frame: Install Pi with the LM Studio plugin and a local model, have Pi scaffold a package agent for one inbox or ticketing task, connect real credentials via .env, then run it unattended with print mode and a polling cron job.
- Why this matters: New playlist item from Bart Slodyczka; queued for transcript-backed review, topic mapping, and a practical learning artifact.

Transcript anchors from this exact video:
- 0:00 / Evidence 1: "There's a really interesting open-source agent harness called pie.dev that's most commonly known for powering the open claw agent. Now, if you've used open claw or Hermes agent, you know that these systems have access to your computer."
- 2:28 / Evidence 2: "locally on our own computer. We've installed Pi and plugged it into LM Studio, and now it's time for step two, to build out our agent. Now, the cool thing about Pi is that when you first install..."
- 3:59 / Evidence 3: "full programmatic control. And then our third option is to build a Pi package, which is a shareable agent. In this case, you can see we have the name, my Pi agent. We have the extensions, the skills,..."
- 5:39 / Evidence 4: "in from somewhere as well, skill.md, response quality, response structure. For initial scaffolding, I think this is really cool, to be honest. It's even giving me tags to use for the ticket and then tools. So I think..."
- 7:52 / Evidence 5: "build the actual agent folder. We've given credentials for our inbox, and our agent's able to read tickets in our inbox and actually generate responses and send them to customers. Now, I would need to go through here..."
- 10:37 / Evidence 6: "identifiable information, defending against prompt injections, or running your agent in the cloud. There's a couple of other concepts that will get your agents to be production ready, ready to actually run in the wild against real customer..."

Video-aware target:
- Prompt lane: Local model/runtime
- Mechanism to extract: Identify why the local setup works or fails for this specific agent task, not whether local models are generally good.
- Artifact to produce: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule.
- Artifact must include: hardware; runtime; model/quantization; endpoint; agent integration; benchmark/fallback

Your task:
1. Use the transcript anchors above as the primary source packet. If you add outside context, label it clearly as outside context and keep it secondary.
2. Create a source-check table with columns: timestamp, claim, transcript support, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable mechanism from the video: Identify why the local setup works or fails for this specific agent task, not whether local models are generally good. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule.
5. Include:
   - a plain-English definition of the core idea
   - a diagram or structured model using this sequence: Task -> Hardware -> Model/quantization -> Runtime endpoint -> Agent tool loop -> Benchmark task -> Fallback
   - answers to these source questions: What machine/runtime is shown? | What task exposes the model limit? | What setup change improves the loop?
   - 3 concrete examples that apply the video idea to real agentic work, such as Ollama or LM Studio coding endpoint; MLX Apple Silicon runner; DGX-backed Hermes session
   - 2 failure modes the video helps prevent, chosen from the transcript evidence and these likely risks: using a local model like ChatGPT; ignoring latency/context limits; no benchmark task
   - a checklist for the next real workflow, focused on: task fit, runtime setup, latency/context, tool loop, fallback
   - one practical exercise with a clear done signal: Choose one real coding task and specify the pass/fail benchmark for a local model.
6. Add a "learning transfer" section: what changes in my workflow tomorrow if I actually learned this?
7. Add a "source check" section that cites which transcript anchor supports each major takeaway.

Quality bar:
- Make this specific to "Qwen 3.6 + Pi Agent: Build Your Own AI Assistant (Full Setup)", not a generic Creative Automation essay.
- Tie each harness element to a transcript anchor that names a tool, state boundary, permission, model behavior, or verification step.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- Avoid these generic drifts: local-model ideology; hardware specs without workflow fit; benchmarks unrelated to the actual task.
- If evidence is weak or missing, stop and say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.