mathiasandClaude Opus 4.8
dcb9ff4a56
The #35 data gate stays at zero because pass-rate only accrues from review/ debug calls *through the routing pod* — Crush, cloud chat, and the local .skills all bypass it. Document the connect → route → flywheel steps, the endpoints (koala:30310/mcp + routing-mcp.d-ma.be), the cold-start behavior (nil pass-rate routes cloud until passes accrue past the 0.90 floor, then local qwen36 activates), the 50-invocation / 2026-07-10 target, and the Berget fallback. Refs #35. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
96 lines
3.9 KiB
Markdown
96 lines
3.9 KiB
Markdown
# Runbook: exercising review/debug traffic to fill the pass-rate dataset
|
||
|
||
**Why this exists:** the routing pod's local-vs-cloud decision is gated on a
|
||
pass-rate history that only accrues from real `review`/`debug` invocations
|
||
**through the pod**. Until the dataset has data, the fast (local) path never
|
||
activates and the core hypothesis (hyperguild #35) can't be validated. This
|
||
runbook is how you spin that flywheel.
|
||
|
||
## The one trap
|
||
|
||
Pass-rate accrues **only** when a skill tool is called via the routing pod's MCP
|
||
endpoint. These look like they should count but **do not**:
|
||
|
||
- **Crush** — talks to LiteLLM directly, bypasses the pod. No log.
|
||
- **claude.ai web / Claude Desktop without the connector** — no log.
|
||
- **Running the local `code-review` / `debug` skills** (`~/dev/.skills`) inline in
|
||
a Claude Code session — those are local skills, not the pod's MCP tools. No log.
|
||
|
||
Only a `tools/call` to the routing pod records a pass/fail.
|
||
|
||
## Endpoints
|
||
|
||
| Purpose | URL | Auth |
|
||
|---------|-----|------|
|
||
| Routing MCP (local, Tailscale) | `http://koala:30310/mcp` | Bearer `ROUTING_MCP_TOKEN` |
|
||
| Routing MCP (remote) | `https://routing-mcp.d-ma.be/mcp` | OAuth via `auth.d-ma.be` (audience `claude-ai`) |
|
||
| Pass-rate readout | `http://koala:30330/pass-rate?skill=<name>` | none (read-only) |
|
||
|
||
Tools advertised: **`review`**, **`debug`** (the two the #35 gate measures),
|
||
plus `session_log`, `retrospective`, `trainer`.
|
||
|
||
## Step 1 — connect the routing pod as an MCP server
|
||
|
||
**Local** (needs the bearer token; keep it out of argv via 1Password):
|
||
|
||
```bash
|
||
op run --env-file ~/.op-env -- \
|
||
claude mcp add routing --transport http http://koala:30310/mcp \
|
||
--header "Authorization: Bearer $ROUTING_MCP_TOKEN"
|
||
```
|
||
|
||
**Remote** (claude.ai / Claude Desktop): add a custom connector pointing at
|
||
`https://routing-mcp.d-ma.be/mcp`; it completes OAuth against `auth.d-ma.be`,
|
||
no static token.
|
||
|
||
Verify: a `tools/list` should return `review`, `debug`, `session_log`,
|
||
`retrospective`, `trainer`.
|
||
|
||
## Step 2 — route real work through it
|
||
|
||
In normal sessions, invoke the pod's tools instead of reviewing/debugging inline:
|
||
|
||
- *"Use the **routing** `review` tool on this diff."*
|
||
- *"**debug** this failure through the routing pod."*
|
||
|
||
Each call logs an outcome to ingestion → `/pass-rate` ticks up.
|
||
|
||
## Step 3 — the flywheel (cold-start behavior is the gate)
|
||
|
||
The router's cold start **is** the data gate — same loop:
|
||
|
||
1. With pass-rate `nil` (cold), the policy routes to the **thinking / cloud**
|
||
model — it won't trust the local fast model without history.
|
||
2. As `review` / `debug` accumulate **passes**, pass-rate climbs past the floor
|
||
(`HYPERGUILD_ROUTE_LOCAL_FLOOR=0.90`) and the router starts sending those
|
||
skills to the **local fast tier**.
|
||
3. So early calls "pay in" on cloud to build the record; then the local tier
|
||
activates automatically. Exercising the traffic is what spins it up.
|
||
|
||
The fast tier is now **`koala/qwen36-35b-a3b`** (Qwen3.6-35B-A3B MTP), promoted
|
||
2026-06-29 after clearing the tool-call gate (infra `c66a195`,
|
||
`HYPERGUILD_FAST_MODEL`).
|
||
|
||
## Target & verification
|
||
|
||
- **50 logged invocations** across `review` + `debug` within the 14-day window
|
||
(opened 2026-06-26, **kill-date 2026-07-10**) → ~4–5 calls/day.
|
||
- Check progress anytime:
|
||
|
||
```bash
|
||
curl -s "http://koala:30330/pass-rate?skill=review"
|
||
curl -s "http://koala:30330/pass-rate?skill=debug"
|
||
```
|
||
|
||
- If ~4–5/day isn't realistic alongside Crush, that is **not** a failure — per
|
||
#35 deliverable #1 it's the signal hyperguild isn't on the work critical path,
|
||
and the pre-decided **Berget fallback** (`gpt-oss-120b` / `qwen3-32b`) carries
|
||
the fast tier instead.
|
||
|
||
## Refs
|
||
|
||
- hyperguild #35 — the validation issue (data gate = deliverable #1)
|
||
- `docs/multi-model-routing.md` — routing policy
|
||
- brain: `wiki/homelab/hypotheses/qwen36-35b-a3b-fast-model-experiment-2026-05-28.md`
|
||
- infra `c66a195` — qwen36 promotion; `models.yml` / `llama-swap-configmap.yaml`
|