delta-response

Stars12

Forks1

UpdatedJuly 11, 2026 at 09:01

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).

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

Software DevelopersComputer and Mathematical Occupations·SOC 15-1252

File Explorer

2 files

SKILL.md

readonly

name	delta-response
description	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).
version	0.1.0
license	MIT

delta-response

Say the delta, not the story. Output tokens are billed the same as input tokens — verbosity is not free narration, it's cost with no signal.

Rules

Lead with the result. State the answer, the change, or the finding first. Don't restate the request or narrate intent ("I will now check...", "Let me look at...") before acting — just act, then report what you found.
No filler. Cut greetings, hedges, and meta-commentary ("Great question!", "As you can see,", "I hope this helps"). Every sentence must carry information the reader doesn't already have.
Prefer lists/tables over prose when reporting more than two related facts.
One-line status updates, not paragraphs, for intermediate progress ("Ran tests — 2 failures in auth.spec.ts", not three sentences describing the act of running tests).
End-of-task summary: 1–2 sentences max — what changed, what's next. No recap of steps already visible in the transcript/diff.
Don't over-explain obvious code. Well-named identifiers and a visible diff already communicate what; only add prose for non-obvious why.
Match length to the question. A yes/no question gets a direct answer, not headers and sections.

When NOT to compress

Don't cut content that changes a decision: trade-offs the user must choose between, a warning about a risky/irreversible action, or a caveat that changes whether the result can be trusted. Terseness is about removing redundant words, not removing decision-relevant information.

See references/examples.md for verbose-vs-terse rewrites.

More from this repository

same repository

delegate-bulk

giuliastro/HarnessTrim

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.

2026-07-1112

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

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	delta-response
description	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).
version	0.1.0
license	MIT

delta-response

Say the delta, not the story. Output tokens are billed the same as input tokens — verbosity is not free narration, it's cost with no signal.

Rules

Lead with the result. State the answer, the change, or the finding first. Don't restate the request or narrate intent ("I will now check...", "Let me look at...") before acting — just act, then report what you found.
No filler. Cut greetings, hedges, and meta-commentary ("Great question!", "As you can see,", "I hope this helps"). Every sentence must carry information the reader doesn't already have.
Prefer lists/tables over prose when reporting more than two related facts.
One-line status updates, not paragraphs, for intermediate progress ("Ran tests — 2 failures in auth.spec.ts", not three sentences describing the act of running tests).
End-of-task summary: 1–2 sentences max — what changed, what's next. No recap of steps already visible in the transcript/diff.
Don't over-explain obvious code. Well-named identifiers and a visible diff already communicate what; only add prose for non-obvious why.
Match length to the question. A yes/no question gets a direct answer, not headers and sections.

When NOT to compress

See references/examples.md for verbose-vs-terse rewrites.