AGENT SKILL

tighten-your-output

Re-anchors terseness discipline (markdown in fewer words without semantic loss, code in fewer lines when readability holds), then audits the work in flight for avoidable verbosity and tightens it. Use when: 'tighten your output', 'tighten this', 'too verbose', 'say it in fewer words', 'this is bloated', 'trim the code', 'simpler form', 'cut the wordiness', 'be more concise', or at conversation start on prose- or code-heavy work.

License 许可
MIT
最近更新
2026-11-15
Stars
★ 30
Forks
7

01 前言 Front Matter

metadata
{ "summary": "Tighten prose and code. Fewer words, no semantic loss", "workflow-stage": "anytime", "discipline-batch": "core", "discipline-batch-rank": 110 }
description
Re-anchors terseness discipline (markdown in fewer words without semantic loss, code in fewer lines when readability holds), then audits the work in flight for avoidable verbosity and tightens it. Use when: 'tighten your output', 'tighten this', 'too verbose', 'say it in fewer words', 'this is bloated', 'trim the code', 'simpler form', 'cut the wordiness', 'be more concise', or at conversation start on prose- or code-heavy work.
user-invocable
true
disable-model-invocation
false

03 SKILL.md 正文

Tighten your output

A drift corrector for terseness discipline. The shared method (re-anchor, audit the work in flight, correct forward, report) lives in ${CLAUDE_PLUGIN_ROOT}/context/re-anchor-audit-correct.md. Read it; this file adds only what is specific to terseness discipline.

The discipline this re-anchors

Say it in fewer words or lines when nothing of value is lost. Two surfaces, resolved separately per the method doc's ladder.

Code, a standards convention owns this

The source of truth is the consuming organization's simpler-code convention: prefer the smaller form when it costs no correctness, readability, test coverage, observability, or convention conformance. Read it where it lives; re-anchor its two load-bearing parts rather than restating them:

  • its named failure modes, the canonical smells a reduction must not introduce (speculative generality, the wrong abstraction, YAGNI);
  • its constraints never traded for line count. Clarity, test coverage, error handling at boundaries, established conventions, and observability survive every cut.

When the consuming project declares no such convention, re-anchor that same shape as the portable baseline: fewer lines is the default only when none of those constraints is spent to get there.

Markdown. A flagged standards gap with a routing owner

Say the same thing in fewer words with no semantic loss. Drop filler, hedging, and restatement, never a directive, qualifier, threshold, or example. Tightening selects what to include; it does not compress prose into fragments, abbreviations, arrow chains, or jargon, and it keeps complete sentences a reader who did not watch the work can follow. When brevity and readability conflict, readability wins. Prose terseness usually has no dedicated standards convention: when the consuming project's standards source declares one, route through it; when it does not, that is a flagged gap (a candidate upstream standards addition), not license to invent a rubric in this skill. Route reader-facing prose and batch prose work per the Routing section below.

Audit. What to look for

Name concrete, located findings (per the method doc's step 2, self-audit; on load-bearing work a fresh-context subagent confirms a clean result):

  • prose padded with filler, hedging, or a restatement of what an adjacent sentence already said;
  • code carrying speculative parameters, dead branches, or hand-rolled boilerplate a built-in would replace;
  • a long form kept where a shorter one loses nothing of value.

Correct each forward now, but only where the reduction is free: tighten the prose without dropping a directive/qualifier/threshold/example, and shrink the code without spending clarity, tests, error handling, conventions, or observability. When a cut would cost any of those, leave the longer form and say why.

Resolve the verification level once with bash "${CLAUDE_PLUGIN_ROOT}/lib/read-verification.sh" --plugin discipline --skill tighten-your-output --file docs/conventions/discipline.yaml and report its line (levels): light and off skip the method doc's step 2 and step 3 subagents, as it says, and the final reply then carries the line verification: skipped (<level>) verbatim, such as verification: skipped (light).

Routing, where batch and proactive work live

  • Prose a person reads (a ticket comment, a pull-request body, a doc, a status update) → /writing:be-concise when the writing plugin is installed. It owns the point-first, scannable shape those readers need, which this corrector does not. Without that plugin, reduce against the markdown discipline above and say the point-first shape went unapplied.
  • Batch prose remediation → the docs-hygiene compress capability when that plugin is installed (its semantic-diff safety net is the guardrail for large prose cuts); otherwise run that same semantic-diff discipline as an explicit in-thread pass rather than hand-rolling a rubric here.
  • Batch code reduction → the code-tidying batch-simplify capability over a change set when that plugin is installed; otherwise reduce in-thread against the simpler-code convention, spending no clarity, test, error handling, convention, or observability to do it.
  • Proactive code-side enforcement is NOT here. It lives in review and design gates that cite the simpler-code convention; this corrector re-anchors the discipline and corrects the work in flight.

Trigger ownership. One owner per brevity phrase

Two standing brevity postures exist, so the vocabularies are split rather than shared. The test is what the phrase names: the output in front of us, or an artifact a reader will open. A phrase that names neither is a register directive, and its default target is the conversation, so it stays here. Every phrase has exactly one owner, and neither description claims a phrase this table gives the other.

Phrase Owner Why
'tighten your output' this skill Names the assistant's own output
'tighten this' this skill Unbound target, so the thing in flight
'too verbose' this skill A verdict on what was just produced
'say it in fewer words' this skill Re-saying inside the conversation
'this is bloated' this skill Unbound target, most often code
'trim the code' this skill Code terseness
'simpler form' this skill Code terseness
'cut the wordiness' this skill A register directive, no artifact named
'be more concise' this skill A register directive, no artifact named
'write this for the PO' writing:be-concise Names the reader
'shorten this ticket' writing:be-concise Names the artifact
'make this scannable' writing:be-concise A property only a human reader needs
'bottom line first' writing:be-concise A property only a human reader needs
'nobody will read this' writing:be-concise Names the reader
'rewrite this PR description for reviewers' writing:be-concise Names the artifact and its readers

Once either skill is loaded, route by the target in front of you, not by the phrase that reached it; the sections above say where.

Gotchas

  • Terseness is not duplication. Removing a second copy of a fact is /discipline:point-dont-copy; removing wasted words from a single copy is this skill.
  • The failure mode is over-cutting: a "concise" edit that quietly drops a qualifier or a boundary case has traded meaning for length. Judge every cut by what it costs, not by how many words it saves.

04 同仓库更多 Skills

melodic-software
melodic-software/claude-code-plugins

track

Backlog reads through the bound tracker: search open and closed items, view one item with its comments and state history. Use when another skill searches or reads tracker items.

★ 30
melodic-software
melodic-software/claude-code-plugins

reviewer-media

Reviewer media for one pull request: decides whether it needs a narrated walkthrough, builds the storyboard from /review:explain-change's digest, has /explainer-video:produce render it and hands the manifest to /source-control:pull-request media. Use when asked for 'reviewer media', 'a walkthrough video for this PR', 'a produced video of this change', or when pull-request ready or monitor hands a PR over.

★ 30
melodic-software
melodic-software/claude-code-plugins

route

Model and effort per agent role (orchestrator, worker, verifier, retrieval) for a multi-agent run, resolved through the bundled defaults and the user, team and personal override config layers into fenced JSON with single-agent and fan-out variants and the layer behind each value. Use when asked 'which model for the subagents', 'route the roles' or 'model routing', or before a skill passes a role's model on a spawn.

★ 30
melodic-software
melodic-software/claude-code-plugins

route

Model and effort per agent role (orchestrator, worker, verifier, retrieval) for a multi-agent run, resolved through the bundled defaults and the user, team and personal override config layers into fenced JSON with single-agent and fan-out variants and the layer behind each value. Use when asked 'which model for the subagents', 'route the roles' or 'model routing', or before a skill passes a role's model on a spawn.

★ 30
melodic-software
melodic-software/claude-code-plugins

principles

Module-design vocabulary for architecture findings, and the checks built on it. Use when: 'what is a deep module', 'what is a seam', 'deep vs shallow module', 'what is the deletion test', 'leverage and locality', 'module design vocabulary', 'which word do we use for X in architecture findings'. Scanning a codebase for friction is /architecture:improve.

★ 30
melodic-software
melodic-software/claude-code-plugins

write

Structured five-field bug report (title, steps to reproduce, expected vs actual, justified severity, suggested fix location) from an informal description, read-only. Use when the user names a defect they observed, or asks for one written up ('report a bug', 'file a bug', 'write this up as a bug').

★ 30