summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorJesse Vincent <jesse@fsck.com>2026-08-19 17:32:55 +0000
committerJesse Vincent <jesse@fsck.com>2026-08-19 17:32:55 +0000
commita56a34365aacd24d16be845f679d2d14f39ae4dd (patch)
treefe06c08b2ed785c8d575766ac7e64c5b941f6078
parentef0ca09658ae9993add7f7c0ea877b90b8166c61 (diff)
feat(sdd): follow a task's declared implementation tier at dispatchskeleton-alternative
The skeleton-first task template carries a Tier field the planner marks deliberately. Model Selection now follows it instead of re-deciding at dispatch, the implementer template's model slot reads it at fill time, and an explicit mechanical ruling overrides the mid-tier contract floor — the planner has already ruled the deliverable fully specified. Task templates without a Tier field are unaffected: every route falls through to the existing Model Selection guidance.
-rw-r--r--skills/subagent-driven-development/SKILL.md9
-rw-r--r--skills/subagent-driven-development/implementer-prompt.md7
2 files changed, 13 insertions, 3 deletions
diff --git a/skills/subagent-driven-development/SKILL.md b/skills/subagent-driven-development/SKILL.md
index 5e565f9..ac324fc 100644
--- a/skills/subagent-driven-development/SKILL.md
+++ b/skills/subagent-driven-development/SKILL.md
@@ -195,6 +195,10 @@ Use the least powerful model that can handle each role to conserve cost and incr
**Mechanical implementation tasks** (isolated functions, clear specs, 1-2 files): use a fast, cheap model. Most implementation tasks are mechanical when the plan is well-specified.
+When a task carries a **Tier:** field, follow it — the planner already
+ruled: mechanical → the cheapest available model; judgment → a standard
+model. Do not re-litigate the tier at dispatch.
+
**Integration and judgment tasks** (multi-file coordination, pattern matching, debugging): use a standard model.
**Architecture and design tasks**: use the most capable available model.
@@ -216,7 +220,10 @@ most expensive — which silently defeats this section.
**Turn count beats token price.** Wall-clock and context cost scale with how
many turns a subagent takes, and the cheapest models routinely take 2-3× the
turns on multi-step work — costing more overall. Use a mid-tier model as the
-floor for reviewers and for implementers working from prose descriptions.
+floor for reviewers and for implementers working from task contracts or
+prose descriptions — unless the task's Tier line says mechanical: the
+planner has already ruled the deliverable fully specified, so treat a
+mechanical-tier contract like spelled-out content.
When the task's plan text contains the complete code to write, the
implementation is transcription plus testing: use the cheapest tier for
that implementer. Single-file mechanical fixes also take the cheapest tier.
diff --git a/skills/subagent-driven-development/implementer-prompt.md b/skills/subagent-driven-development/implementer-prompt.md
index 5c8ecd6..e8f805b 100644
--- a/skills/subagent-driven-development/implementer-prompt.md
+++ b/skills/subagent-driven-development/implementer-prompt.md
@@ -5,8 +5,11 @@ Use this template when dispatching an implementer subagent.
```
Subagent (general-purpose):
description: "Implement Task N: [task name]"
- model: [MODEL — REQUIRED: choose per SKILL.md Model Selection; an omitted
- model silently inherits the session's most expensive one]
+ model: [MODEL — REQUIRED: when the brief carries a Tier line, set from it:
+ mechanical → the cheapest model the subagent tool offers; judgment →
+ a standard mid-tier model. Otherwise choose per SKILL.md Model
+ Selection. An omitted model silently inherits the session's most
+ expensive one]
prompt: |
You are implementing Task N: [task name]