delegate-bulk

Stars12

Forks1

UpdatedJuly 11, 2026 at 10:00

This skill should be used when deciding whether to hand volumetric or noisy work to an isolated subagent (or another harness) instead of doing it in the main context — e.g. reading many files, digesting long logs/docs, wide searches, or bulk mechanical generation. It gives the rule for when isolation saves tokens versus when it just multiplies them. Part of the HarnessTrim token-economy stack.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Source

giuliastro

giuliastro/HarnessTrim

View GitHub Repository View Creator Repositories

Download

Related occupationsSOC

Based on SOC occupation classification

Computer Occupations, All OtherComputer and Mathematical Occupations·SOC 15-1299

SKILL.md

readonly

name	delegate-bulk
description	This skill should be used when deciding whether to hand volumetric or noisy work to an isolated subagent (or another harness) instead of doing it in the main context — e.g. reading many files, digesting long logs/docs, wide searches, or bulk mechanical generation. It gives the rule for when isolation saves tokens versus when it just multiplies them. Part of the HarnessTrim token-economy stack.
version	0.1.0
license	MIT

delegate-bulk

Subagents are not free parallelism — each carries its own context and bills its own tokens. The right move is surgical isolation of noise, not "multi-agent everything."

Delegate to an isolated subagent when

A task will read a large volume the main thread doesn't need to retain — many files, long logs, big docs, wide search results — and you only need the conclusion back, not the raw bytes.
The work is noisy but self-contained: exploring a subsystem, triaging a failing test suite, digesting API docs. Isolate it so its raw output never pollutes the main context.
You need several independent investigations at once and each one's detail is irrelevant to the others.

The token win comes from the subagent returning a short structured answer while its expensive reading stays in a context you throw away.

Do NOT delegate when

The task is small enough that spawning a context costs more than it saves.
The main thread needs the full detail anyway (delegating then re-reading is pure overhead).
You'd spawn many agents that each re-load the same large context — that multiplies tokens, the opposite of the goal.

Cross-harness delegation

The same logic extends across harnesses: route bulk scaffolding/boilerplate to a cheaper or execution-oriented channel and keep architecture, review, and validation on the main one. Suppress the delegated channel's verbose thinking/output from re-entering the main context. Pair with [[scaffold-fast]] for the delegated bulk work and [[compact-handoff]] for what the subagent returns.

Rule of thumb

Delegate to shrink what the main context must hold, never to duplicate it.

More from this repository

same repository

scaffold-fast

giuliastro/HarnessTrim

This skill should be used for mechanical, low-novelty coding work — generating boilerplate, wiring up a component from an existing pattern, adding a CRUD endpoint like the others, writing repetitive tests, or applying a rote transformation across files. It keeps reasoning effort low and output terse for work whose shape is already decided. Part of the HarnessTrim token-economy stack; the lean-scaffold preset pairs it with low reasoning effort.

2026-07-1112

compact-handoff

giuliastro/HarnessTrim

This skill should be used when a session is being compacted/summarized (context is full or the user runs a compact command), or when writing any handoff summary that a later session or subagent will continue from. Defines what to preserve versus drop so the continuation doesn't waste tokens re-reading files and re-deriving state. Part of the HarnessTrim token-economy stack; the OpenCode adapter injects this guidance automatically via experimental.session.compacting.

2026-07-1112

debug-log-slim

giuliastro/HarnessTrim

This skill should be used before pasting, printing, or otherwise embedding test-runner output, build logs, or CI output into the conversation — whenever a test suite, linter, or build was just run and its raw output is large. Filters logs down to failure/error signal before they enter context. Part of the HarnessTrim token-economy stack (see packages/core/src/reducers/test-output-slim.ts for the reference implementation the OpenCode adapter automates this with).

2026-07-1112

delta-response

giuliastro/HarnessTrim

This skill should be used for every response produced in a coding harness (Claude Code, Codex, OpenCode, Pi) — not only when explicitly requested. Enforces terse, structured, information-dense output instead of narrated verbosity, to cut output-token cost. Part of the HarnessTrim token-economy stack (see delta-response/references/examples.md for before/after samples).

2026-07-1112

review-delta

giuliastro/HarnessTrim

This skill should be used when asked to review code, review a pull request, or review a diff. Focuses the review on problems worth fixing rather than restating or summarizing what changed. Part of the HarnessTrim token-economy stack.

2026-07-1112

name	delegate-bulk
description	This skill should be used when deciding whether to hand volumetric or noisy work to an isolated subagent (or another harness) instead of doing it in the main context — e.g. reading many files, digesting long logs/docs, wide searches, or bulk mechanical generation. It gives the rule for when isolation saves tokens versus when it just multiplies them. Part of the HarnessTrim token-economy stack.
version	0.1.0
license	MIT

delegate-bulk

Subagents are not free parallelism — each carries its own context and bills its own tokens. The right move is surgical isolation of noise, not "multi-agent everything."

Delegate to an isolated subagent when

A task will read a large volume the main thread doesn't need to retain — many files, long logs, big docs, wide search results — and you only need the conclusion back, not the raw bytes.
The work is noisy but self-contained: exploring a subsystem, triaging a failing test suite, digesting API docs. Isolate it so its raw output never pollutes the main context.
You need several independent investigations at once and each one's detail is irrelevant to the others.

The token win comes from the subagent returning a short structured answer while its expensive reading stays in a context you throw away.

Do NOT delegate when

The task is small enough that spawning a context costs more than it saves.
The main thread needs the full detail anyway (delegating then re-reading is pure overhead).
You'd spawn many agents that each re-load the same large context — that multiplies tokens, the opposite of the goal.

Cross-harness delegation

Rule of thumb

Delegate to shrink what the main context must hold, never to duplicate it.