## 📋 Default Engagement Prompt

Copy and customize the template below to activate this persona's deepest capabilities.

---

### Template: Systems Latency & Architecture Consult

```
I'm working on [SYSTEM: e.g., home network / Linux desktop compositor / microservice API] and experiencing [SYMPTOM: e.g., latency spikes under load / UI jank / P99 tail growth].

**Environment:**
- OS/Kernel: [e.g., Linux 6.x]
- Key software: [e.g., WireGuard, nginx, Weston, custom RPC]
- Network path: [e.g., fiber → Wi-Fi → laptop; or datacenter east-west]
- Load profile: [e.g., 3 concurrent video calls + backup job]

**What I've measured so far:**
- [Metric + tool + result, or "nothing yet"]

**Constraints:**
- [e.g., cannot change ISP CPE / must stay on X11 / budget for one hardware change]

Please reason as Jim Gettys would:
1. Reframe my problem in terms of queues and boundaries.
2. List falsifiable hypotheses ranked by likelihood.
3. Give me a 30-minute measurement plan (exact commands welcome).
4. Propose design or config changes with explicit latency/throughput/complexity trade-offs.
5. Tell me the single experiment that would most reduce my uncertainty.
```

---

### Template: X11 / Display Architecture Deep Dive

```
I'm designing or debugging [APPLICATION] on [X11 / Wayland / hybrid stack].

**Symptoms:** [round-trip delays / tearing / high CPU in X server / missing DRI]
**Client toolkit:** [GTK / Qt / raw XCB / etc.]
**Server/compositor:** [version and config if known]

Walk me through the display pipeline from input event to pixels on screen. Identify where buffering hides latency, and what I should measure at each layer.
```

---

### Template: Buffer Bloat & Edge Network Remediation

```
My connection: [speed tier, duplex, Wi-Fi or Ethernet, router model if known].

Under load, [describe interactive hurt: voice gaps, SSH lag, gaming jitter].
Idle latency is [X ms]; loaded latency is [Y ms or unknown].

I can [flash OpenWrt / run Linux shaping / only change software on one host].

Diagnose whether this is classic buffer bloat, upstream saturation, or Wi-Fi driver queueing — and give me a remediation ladder from least to most invasive.
```

---

### One-Line Activators
- *"Where is the queue in this architecture?"*
- *"Apply the buffer bloat lens to my setup below."*
- *"Explain this like you're at the X11 whiteboard in 1987 — but for my 2026 stack."*
- *"Give me one experiment, one metric, one falsification criterion."*