Skip to content

ValentinNikolaev/agent-skills / agent-plugins/skills/humanizer

humanizer

Verified

Humanize existing prose while preserving voice and meaning, including targeted English pattern edits. Not drafting or translation. Use ai-detector for analysis.

references/council.md

View on GitHub
# Council for editing decisions

Use the available `council` workflow in Quick mode for the editing decisions.
If that skill is unavailable, follow the bounded procedure here. A project with
required councils uses its existing advisors and prescribed stages for these
questions, rather than convening a duplicate generic team. Check that those
stages cover edit selection and the exact final text; integrate any missing
question into the owning workflow without bypassing its gates.

1. Give three distinct, independent advisor passes the same small packet:
   original passage, requested operation, relevant voice/task excerpts, meaning
   anchors, protected spans, and the specific AI-sounding features to examine.
   Include keeping the original as an option. Do not present a preferred rewrite
   as the expected answer or expose peer responses before independent passes end.
2. Ask for concrete excerpts, why they create a formulaic impression, the
   smallest effective edit or a reason to keep them, and risks to voice or
   meaning. Compare alternatives at the passage level as well as the sentence
   level; identify the strongest reason an edited version might still read as
   templated. Do not rate AI likelihood by council opinion.
3. Synthesize a candidate. Have one advisor from the initial group critique the
   anonymized recommendations alongside the original and exact candidate. Ask
   what formulaic features remain, where voice or meaning drifted, and what the
   leading choice gets wrong. In review-only mode, assess proposed findings and
   suggestions without rewriting the document or editing its files.
4. Decide from the source and applicable voice rules, resolving objections rather
   than counting votes. Choose the smallest effective edit or leave the passage
   unchanged. For a substantive revision after critique, send the exact revised
   text back to the same advisors for reassessment. Minor corrections need only
   the final preservation check unless the project requires more.

Use at most three subagents total, in one initial wave, and reuse them for review.
Keep tiny edits' packets and answers short. If independent agents are unavailable,
use separated advisor passes in the current model and disclose reduced
independence when material. Agreement is model judgement, not evidence of author
preference, reader perception, or detector performance. Neither a council verdict
nor this skill grants file-edit or publication authority beyond the user's task.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall