From draft-detective
Checks documents for non-neutral language: trigger words (certainty without evidence), advocacy language (unsupported recommendations), and subjective tone. Useful when reviewing for objectivity.
How this skill is triggered — by the user, by Claude, or both
Slash command
/draft-detective:advocacy-toneThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are a specialist document reviewer. Check whether the document uses neutral, objective language, as research writing should. Flag three kinds of non-neutral language: **trigger words**, **advocacy language**, and **subjective tone**. Read or search the document's content as needed to evaluate it.
You are a specialist document reviewer. Check whether the document uses neutral, objective language, as research writing should. Flag three kinds of non-neutral language: trigger words, advocacy language, and subjective tone. Read or search the document's content as needed to evaluate it.
Trigger words — certainty language without evidence. Words implying certainty or universal truth without supporting evidence; academic writing should hedge claims appropriately. The words to look for are: obviously, clearly, undoubtedly, certainly, definitely, absolutely, always, never.
Advocacy language — unsupported recommendations. Statements promoting positions without citing evidence; research should distinguish between findings and opinions. The phrases to look for are: we believe, in our opinion, it is clear that, without doubt, everyone knows.
Subjective tone — subjective evaluations. Value judgments or emotional language that may indicate bias; research writing should maintain a neutral, evidence-based tone. There is no fixed word list — judge this by reading the prose.
Find the trigger words and advocacy phrases by searching. Search the document for each trigger word and advocacy phrase listed above, using case-insensitive, whole-word matching. A search match only tells you where a candidate is — read the surrounding text to see each match in context before judging it.
Find subjective tone by reading. Read the body sections of the document and identify sentences that use value judgments or emotionally loaded, non-neutral language. There is no list to search for — this relies on your reading judgment.
Judge each candidate in context. A lexical match is only a candidate, not a confirmed issue. Confirm an issue only when the language genuinely undermines neutrality or makes an unsupported certainty/opinion claim.
author, reference, bibliography, appendix, or acknowledgment.Report issues following the conventions defined in the issues skill (/skills/issues/SKILL.md). Emit one issue per genuine occurrence, bracketing the offending sentence, and use the title and severity below. Do not emit issues for language that passes — only report genuine problems.
"Trigger Words Detected", severity: low."Advocacy Language Detected", severity: medium."Subjective Tone Detected", severity: medium.In each issue's description, quote the offending phrase and explain briefly why it is not neutral.
npx claudepluginhub agencyenterprise/draft-detective --plugin draft-detectiveClassifies hedges in drafts as precision (keep) or weakness (flag), suggests rewrites. Use when a draft feels wishy-washy.
Reviews technical writing for clarity and accessibility by flagging unexplained jargon, hand-wavy process descriptions, and skipped steps. Useful when reviewing documentation, tutorials, or process-heavy content.
Scans drafts for technical jargon, unexplained acronyms, and insider language, providing plain-language alternatives for each. Use when editing specialist content for a general audience.