Setting Up Pi Subagents Because I Keep Doing 2 Dumb Things

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

Source video:
- Title: Setting Up Pi Subagents Because I Keep Doing 2 Dumb Things
- URL: https://www.youtube.com/watch?v=t70YWb03vm0
- Topic: Agent Architecture
- My current learning frame: Install the Pi sub-agents package, define three minimal agents (scout, oracle, reviewer) in your main prompt, then dictate a background scout-to-oracle chain and verify results come back via intercom while your main context window stays under 10%.
- Why this matters: New playlist item from Eric Michaud; queued for transcript-backed review, topic mapping, and a practical learning artifact.

Transcript anchors from this exact video:
- 0:20 / Evidence 1: "why to install sub agents onto our pie harness. Let's get into it. All right, quick catch up for those of you still using Claude code or Codex or whatever. Pie is an agent harness just like those..."
- 2:17 / Evidence 2: "review, right? So, with that, in the top agent system prompt, I need to tell it exactly who does what, what context those agents need, whether it's a foreground or background task, what needs to be reviewed once..."
- 4:04 / Evidence 3: "different agents that they've got, right? Like um scout, researcher, planner, worker, reviewer, context builder, oracle, delegate, right? So, I think I just want these guys for now. So, this is interesting. It's taking a lot of the..."
- 6:07 / Evidence 4: "if it matches where I'm going with the task today. So, the agent reviewer working in the background. Okay, this is working the way that I wanted it to now. Reviewer says Scout is mostly aligned. Today's video..."
- 8:10 / Evidence 5: "one. Sounds great. I don't want to go down this rabbit hole quite yet. I'm just going to see one more thing here, too. Is there a way to assign different models and thinking levels to these different..."
- 10:38 / Evidence 6: "main system prompt now, you know, like the one for my baseline pi agent to trim it back and make sure that we're not giving all this extra context to every agent. I'd like you to use the..."
- 12:12 / Evidence 7: "context on different angles that I want to look up for a YouTube video and then turn that into a brief for me, right? Like these sub-agents are going to be super super good in making sure things..."

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, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable claims from the video. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A one-page agent harness map with tool boundaries and proof signals.
5. Include:
   - a plain-English definition of the core idea
   - a diagram or structured model using this sequence: Intent -> Model -> Harness -> Tools -> Verifier -> Artifact
   - 3 concrete examples that apply the video idea to real agentic work
   - 2 failure modes the video helps prevent
   - a checklist I can use the next time I run Codex or Claude
   - one practical exercise with a clear done signal
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 "Setting Up Pi Subagents Because I Keep Doing 2 Dumb Things", not a generic Agent Architecture essay.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- If evidence is weak, say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.