From 2ca2c6c42eda564ca6a5699a8c4669b63d730108 Mon Sep 17 00:00:00 2001 From: damocles Date: Mon, 5 Oct 2026 01:56:34 +0200 Subject: [PATCH 1/4] subagent: tell agents they can liberally spawn haiku, even at high effort --- .../base/skills/claude-subagents/SKILL.md | 26 ++++++++++++++++--- hive-subagent-mcp/src/mcp.rs | 12 ++++++--- 2 files changed, 31 insertions(+), 7 deletions(-) diff --git a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md index 352ebf9f5..bf18ff6a7 100644 --- a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md +++ b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md @@ -52,13 +52,33 @@ Model choice, prompt hygiene, splitting big batches, verify-then-report — everything else in this skill — applies exactly the same whether you're calling the tool or thinking through the recipe by hand. +## Model choice + +Pick the cheapest model that can do the job, and lean toward the cheap end +more readily than you might for your own work — a subagent is throwaway +compute, not a peer you're rationing. Haiku in particular is cheap enough +that you don't need to be precious about spinning one up: for a mechanical, +well-scoped batch (the shape this skill is for), reach for it liberally +rather than defaulting to a bigger model "to be safe." Haiku also stays +cheap even at `effort: high` — unlike a frontier model, where effort and +cost scale together, so there's no need to cap Haiku's effort the way the +Effort section below recommends for cost reasons; raise it freely if the +task benefits from the extra reasoning. + +This isn't "always pick the smallest model regardless of task" — a batch +that actually needs judgement calls still wants a stronger model (see +Pitfalls below). It's "don't under-use Haiku out of habit-driven caution +when the task is mechanical enough for the skill to apply at all." + ## Effort An omitted `effort` defaults to `medium` here — cheaper than claude's own model default (`high` on most models), a deliberate cost-conscious choice -for subagent work specifically, same spirit as "cheaper-than-you" model -choice above. Raise it explicitly when the task is complex enough to -actually need deeper reasoning, not as a reflex. +for subagent work specifically, same spirit as the model-choice guidance +above. Raise it explicitly when the task is complex enough to actually +need deeper reasoning, not as a reflex. **Exception: on Haiku, cost doesn't +scale with effort the way it does on a frontier model** — see Model choice +above, don't apply this same-default-medium caution to a Haiku subagent. Anthropic's own levels (per current Claude Code docs — names/availability are model-dependent, check before relying on an exact list): `low`, diff --git a/hive-subagent-mcp/src/mcp.rs b/hive-subagent-mcp/src/mcp.rs index cbe70b189..7a4b84547 100644 --- a/hive-subagent-mcp/src/mcp.rs +++ b/hive-subagent-mcp/src/mcp.rs @@ -44,9 +44,11 @@ struct StartArgs { /// session under that name is done — rejected while one under the same /// name is still running. name: String, - /// Which model the subagent's own session runs. Omit for claude's own - /// default. The `base:claude-subagents` skill's "cheaper-than-you" - /// guidance still applies here. + /// Which model the subagent's own session runs (`--model`). Omit for + /// claude's own default. The `base:claude-subagents` skill's "Model + /// choice" guidance still applies here — pick the cheapest model that + /// can do the job, and don't be stingy about reaching for Haiku on a + /// well-scoped mechanical batch. #[serde(default)] model: Option, /// Which reasoning effort level the subagent's own session runs at @@ -55,7 +57,9 @@ struct StartArgs { /// models). Independent of `model`, so a cheap model at high effort or /// an expensive one at low effort are both valid combinations, not /// just the two extremes. See the `base:claude-subagents` skill for - /// Anthropic's own guidance on choosing between levels. + /// Anthropic's own guidance on choosing between levels — note that on + /// Haiku specifically, cost doesn't scale with effort the way it does + /// on a frontier model, so there's no cost reason to cap it at `medium`. #[serde(default)] effort: Option, /// Path to a file holding the subagent's actual task instructions. A From 69de09e0f906f4f055201c4f4ec7e52c2c1a0ec0 Mon Sep 17 00:00:00 2001 From: damocles Date: Mon, 5 Oct 2026 02:00:25 +0200 Subject: [PATCH 2/4] subagent: drop the unsupported 'raise effort for extra reasoning' claim on haiku --- .../plugins/base/skills/claude-subagents/SKILL.md | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md index bf18ff6a7..774625a56 100644 --- a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md +++ b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md @@ -60,10 +60,14 @@ compute, not a peer you're rationing. Haiku in particular is cheap enough that you don't need to be precious about spinning one up: for a mechanical, well-scoped batch (the shape this skill is for), reach for it liberally rather than defaulting to a bigger model "to be safe." Haiku also stays -cheap even at `effort: high` — unlike a frontier model, where effort and -cost scale together, so there's no need to cap Haiku's effort the way the -Effort section below recommends for cost reasons; raise it freely if the -task benefits from the extra reasoning. +cheap regardless of `effort` level — unlike a frontier model, cost on +Haiku doesn't scale with effort because effort isn't a real lever on it +at all (measured: `thinking_tokens` doesn't visibly move between `low` +and `max`). So there's no cost reason to cap Haiku's effort the way the +Effort section below recommends for frontier models — but don't expect +raising it to buy extra reasoning either; a Haiku batch that genuinely +needs more judgement wants a stronger model instead (see Pitfalls below), +not a higher effort level on the same model. This isn't "always pick the smallest model regardless of task" — a batch that actually needs judgement calls still wants a stronger model (see From a1f1f04657fb65f6a8f07ac09b8cf7cee3dc6bbe Mon Sep 17 00:00:00 2001 From: damocles Date: Sun, 1 Nov 2026 12:24:40 +0100 Subject: [PATCH 3/4] subagent: haiku 5-5 is near free up to high effort on a small context, but max costs about what sonnet does --- .../base/skills/claude-subagents/SKILL.md | 22 +++++++++---------- hive-subagent-mcp/src/mcp.rs | 7 +++--- 2 files changed, 14 insertions(+), 15 deletions(-) diff --git a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md index 774625a56..222cc0886 100644 --- a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md +++ b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md @@ -59,15 +59,13 @@ more readily than you might for your own work — a subagent is throwaway compute, not a peer you're rationing. Haiku in particular is cheap enough that you don't need to be precious about spinning one up: for a mechanical, well-scoped batch (the shape this skill is for), reach for it liberally -rather than defaulting to a bigger model "to be safe." Haiku also stays -cheap regardless of `effort` level — unlike a frontier model, cost on -Haiku doesn't scale with effort because effort isn't a real lever on it -at all (measured: `thinking_tokens` doesn't visibly move between `low` -and `max`). So there's no cost reason to cap Haiku's effort the way the -Effort section below recommends for frontier models — but don't expect -raising it to buy extra reasoning either; a Haiku batch that genuinely -needs more judgement wants a stronger model instead (see Pitfalls below), -not a higher effort level on the same model. +rather than defaulting to a bigger model "to be safe." On a small context +Haiku costs next to nothing up to `high` effort, so there's no cost reason +to cap its effort at `medium` by reflex. Effort is a real lever on Haiku +though: at `max` it behaves like a slightly weaker Sonnet at about Sonnet's +cost, so `max` on Haiku buys nothing over picking Sonnet. Stay at `high` or +below, and move to a stronger model when the batch needs more judgement +(see Pitfalls below). This isn't "always pick the smallest model regardless of task" — a batch that actually needs judgement calls still wants a stronger model (see @@ -80,9 +78,9 @@ An omitted `effort` defaults to `medium` here — cheaper than claude's own model default (`high` on most models), a deliberate cost-conscious choice for subagent work specifically, same spirit as the model-choice guidance above. Raise it explicitly when the task is complex enough to actually -need deeper reasoning, not as a reflex. **Exception: on Haiku, cost doesn't -scale with effort the way it does on a frontier model** — see Model choice -above, don't apply this same-default-medium caution to a Haiku subagent. +need deeper reasoning, not as a reflex. **Exception: on Haiku with a small +context, cost stays near zero up to `high`.** See Model choice above: skip this +default-medium caution there, but keep `max` for a stronger model. Anthropic's own levels (per current Claude Code docs — names/availability are model-dependent, check before relying on an exact list): `low`, diff --git a/hive-subagent-mcp/src/mcp.rs b/hive-subagent-mcp/src/mcp.rs index 7a4b84547..09e745399 100644 --- a/hive-subagent-mcp/src/mcp.rs +++ b/hive-subagent-mcp/src/mcp.rs @@ -57,9 +57,10 @@ struct StartArgs { /// models). Independent of `model`, so a cheap model at high effort or /// an expensive one at low effort are both valid combinations, not /// just the two extremes. See the `base:claude-subagents` skill for - /// Anthropic's own guidance on choosing between levels — note that on - /// Haiku specifically, cost doesn't scale with effort the way it does - /// on a frontier model, so there's no cost reason to cap it at `medium`. + /// Anthropic's own guidance on choosing between levels. On Haiku with a + /// small context, effort up to `high` costs next to nothing, so there's + /// no cost reason to cap it at `medium`; at `max` it costs about what + /// Sonnet does. #[serde(default)] effort: Option, /// Path to a file holding the subagent's actual task instructions. A From b2e3b0519854f97ca4467091bb272a20805ad2ae Mon Sep 17 00:00:00 2001 From: damocles Date: Sun, 1 Nov 2026 12:29:55 +0100 Subject: [PATCH 4/4] subagent: compare haiku max to sonnet on low, and warn that sonnet defaults to high --- .../plugins/base/skills/claude-subagents/SKILL.md | 9 +++++---- hive-subagent-mcp/src/mcp.rs | 2 +- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md index 222cc0886..d3d667f27 100644 --- a/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md +++ b/claude-plugins/plugins/base/skills/claude-subagents/SKILL.md @@ -62,10 +62,11 @@ well-scoped batch (the shape this skill is for), reach for it liberally rather than defaulting to a bigger model "to be safe." On a small context Haiku costs next to nothing up to `high` effort, so there's no cost reason to cap its effort at `medium` by reflex. Effort is a real lever on Haiku -though: at `max` it behaves like a slightly weaker Sonnet at about Sonnet's -cost, so `max` on Haiku buys nothing over picking Sonnet. Stay at `high` or -below, and move to a stronger model when the batch needs more judgement -(see Pitfalls below). +though: at `max` it behaves like a slightly weaker Sonnet at about the cost +of Sonnet on `low`, so `max` on Haiku buys nothing over Sonnet on `low`. +Stay at `high` or below, and move to a stronger model when the batch needs +more judgement (see Pitfalls below). When you do pick Sonnet, pass `low` +yourself: its default effort is `high`, a big price bump. This isn't "always pick the smallest model regardless of task" — a batch that actually needs judgement calls still wants a stronger model (see diff --git a/hive-subagent-mcp/src/mcp.rs b/hive-subagent-mcp/src/mcp.rs index 09e745399..76f528fab 100644 --- a/hive-subagent-mcp/src/mcp.rs +++ b/hive-subagent-mcp/src/mcp.rs @@ -60,7 +60,7 @@ struct StartArgs { /// Anthropic's own guidance on choosing between levels. On Haiku with a /// small context, effort up to `high` costs next to nothing, so there's /// no cost reason to cap it at `medium`; at `max` it costs about what - /// Sonnet does. + /// Sonnet does on `low` (Sonnet's own default is `high`, a big price bump). #[serde(default)] effort: Option, /// Path to a file holding the subagent's actual task instructions. A