71757faa90
* Sync agent pipeline artifacts to 0.6.0. Add model routing defaults, L2 role updates, convoy telemetry, and manifest tracking without touching unrelated in-progress work. * Record tcg-vault interactive sync (kept local L1/L3 customizations). Update last_synced_at after reviewing behind/conflict artifacts; no overwrites applied. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
34 lines
1.7 KiB
Text
34 lines
1.7 KiB
Text
---
|
|
description: Model tier defaults for the agent pipeline — Opus for planning only; fast models for build and audit
|
|
alwaysApply: true
|
|
---
|
|
|
|
# Model routing (cost-aware pipeline)
|
|
|
|
Cursor subagent roles in `.cursor/agents/role-*.md` declare a `model:` in frontmatter. **Invoke roles from the Agents dropdown** (or a fresh chat pinned to that role) so the configured model applies. Parent chats on Opus that spawn Task subagents may override role models — see `docs/model-routing-policy.md`.
|
|
|
|
## Default session model
|
|
|
|
- **Parent chat (ad hoc work):** `auto` or `composer-2.5-fast`.
|
|
- **Escalate to Opus** (`claude-4.6-opus-high-thinking`) only for conductor, architect, or after two failed implementer attempts on the same brief.
|
|
- **Never use Opus 4.7 / xhigh** for routine pipeline work unless leadership explicitly approves a spike.
|
|
|
|
## Per-role tiers (canonical)
|
|
|
|
| Tier | Roles | Model |
|
|
| --- | --- | --- |
|
|
| Premium | conductor, architect | `claude-4.6-opus-high-thinking` |
|
|
| Fast | ia-architect, ux-reviewer, implementer, reviewer, design-system-auditor, a11y-auditor | `composer-2.5-fast` |
|
|
| Auto | doc-writer | `auto` |
|
|
|
|
Read `model_policy:` from the convoy frontmatter when present; it overrides these defaults for that convoy.
|
|
|
|
## Context discipline (biggest cost lever)
|
|
|
|
- Implementers and auditors: read **only** the brief + files in `files:` — not the whole convoy, not `prisma/schema.prisma` unless listed.
|
|
- Audits: diff + brief only; no repo-wide grep sweeps.
|
|
- Stop and ask before any single chat exceeds ~5M tokens; start a fresh scoped chat instead.
|
|
|
|
## Metrics
|
|
|
|
After each role, log `model=` and `model_tier=` via `scripts/log-convoy-event.sh` (see each role's Metrics section).
|