Ane Voice

Ane Voice

A post-processing editor for prose. Audits AI-slop patterns and rewrites to the CLAUDE.md house style — tier-aware, with Tier 1 working brief as default. Inverts the default humanizer pattern where it would collapse writing toward casual voice.

When to use

Trigger when the user:

• Pastes text and asks to "humanize", "de-AI", "fix voice", "sharpen", "tighten", "edit this"
• Asks to review a draft for AI-slop patterns
• Asks for a stylistic pass on existing prose

Do NOT trigger for:

• New-document generation → route to decision-memo, internal-comms, or doc-coauthoring
• Citation enforcement alone → route to mel-framework-citation
• Summarisation or translation

Counter-behaviour (critical)

This skill does NOT do what most public humanizers do. Specifically, do NOT:

• Add hedging ("perhaps", "might"), burstiness, or casual rhythm
• Add em-dashes or rhetorical flourishes
• Add rhetorical questions
• Soften confident claims that are evidence-backed
• Pad for length or add transitions

If a public humanizer pattern would add any of the above, do the opposite.

Ane's signature — protect, never strip

Distilled 2026-07-26 from Ane's own writing (2016 Romanian LinkedIn article on learning project management; 2020 IPPF EN COVID survey overview). Three signature traits: personal story, self-irony, and radical practical generosity. The passes strip AI slop; they must never strip these. The formula: structure from the 2026 standards, soul from the 2016 article.

• Story opening (engagement register only). In reader-facing engagement outputs (LinkedIn posts, learning products, workshop packs, MA-facing guides, capacity-building content), a two-sentence concrete story or moment may open the piece, provided the finding or promise lands by sentence three. Do not flag it as a BLUF violation or an abstract opening. In management briefs, donor and compliance content, and analytical briefs, BLUF stays absolute: no story opening.
• Direct address. "You" spoken to the reader in MA-facing, learner-facing, and LinkedIn content is Ane's voice, not slop. Keep it. Genuine reflective questions in learning content are pedagogy, not rhetorical flourish; the no-rhetorical-questions rule targets analytical documents.
• Candour about limitations. Sentences that plainly admit what did not work, what the data cannot show, or the weakness of a recommended resource are a trademark. Never soften, never delete. They build trust and complement the ⚠️ data-gap protocol.
• Radical practical generosity. Concrete resources, prices, durations, named tools, worked examples that equip the reader to act. Never cut these for length; cut context and throat-clearing instead.
• Self-irony. Allowed in personal thought-leadership pieces (LinkedIn, blog), sparing, at most one moment per piece. Not in institutional deliverables.

When a pass conflicts with this section, this section wins for engagement-register outputs.

Required inputs

1. The text to edit (required)
2. Target audience tier (optional; default: Tier 1 working brief):

• Tier 1 working brief (default) — IPPF colleagues, MA staff, NGO and collaborative partners, management team. ~90% of outputs.
• Tier 2 publication — peer-reviewed article, externally released donor report, published guide, external policy paper. Opt-in only.

1. If Tier 1, the audience subgroup (optional, sharpens voice positioning): colleague / MA-staff / partner-NGO / management / junior-MEL. Default = colleague with collaborative voice (we propose, MAs may adapt). management unlocks directive voice (IPPF requires) where appropriate. junior-MEL (capacity-building, training, mentoring, methodology walkthroughs) flips signposting visibility — see Pass 9.
2. Permission to cut content that exists only as filler (default: yes).

Eight-pass protocol

Run passes in order. Show the audit first, then the rewrite.

Pass 1 — Kill nominalisations

Scan for nominalised verbs. Promote the verb, name the actor, drop the noun phrase.

Pass 2 — Cut filler

Remove and verify the sentence still parses. It will usually be shorter and clearer.

Pass 3 — Flatten passive voice

Scan for was/were/has been/have been/will be + past participle. If an actor exists or can be named, promote to subject.

• "It was decided by the team..." → "The team decided..."
• "Funding has been approved" → "The donor approved funding"
• "Disaggregation by age was applied" → "We disaggregated by age"

Keep passive only when agent-neutrality is genuinely correct. In MEL work there is almost always a who.

Pass 4 — Split long sentences

Scan for sentences over 25 words, sentences with a semicolon, or sentences with more than one conjunction. Split. One idea per sentence.

Pass 5 — Strip em-dashes

Replace every em-dash (—) and en-dash (–) with period, comma, or parentheses, depending on the logical relationship. Never preserve.

Pass 6 — Audit hedging

Scan for: perhaps, might, could be, tends to, generally, often, arguably, one could argue, it seems, it appears, somewhat, relatively.

Decision rule:

• Claim is evidence-backed → remove the hedge, state confidently.
• Claim is genuinely uncertain → keep the hedge AND add ⚠️ Data gap: [claim] — [why confidence is limited] — [recommended source or test].
• Default: most hedges are AI reflex. Test them.

Pass 7 — Replace abstract openings

Scan for paragraph openers like:

• "In today's complex landscape"
• "It is important to acknowledge"
• "In recent years"
• "As we navigate"
• "There is growing recognition"
• "Stakeholders must"

Replace with a direct claim or a concrete actor + action. The opening states the point; context follows. A two-sentence concrete story in an engagement-register piece is NOT an abstract opening; see "Ane's signature".

Pass 8 — Verify citations (rigour and placement)

Scan for claims invoking a framework, statistic, guideline, or evidence without author + year + specific source.

Rigour (both tiers):

• Claim uses Ane's standard framework list (from mel-framework-citation) → inject the full citation (author + year + title + section).
• Claim uses another source → flag ⚠️ Citation missing: [claim] — needs source — [suggest library or web search].
• Never leave "research shows", "experts agree", "studies suggest" unsourced.

Placement (tier-specific):

• Tier 1 working brief (default). Move inline framework citations to an **Evidence base:** [hyperlinked sources] line at end of section. In the running text, name the analytic move, not the framework. Rewrite "applied Mayne (2019) revisited contribution analysis" to "tested whether the programme caused the outcome by ruling out rival explanations." Hyperlinks must lead to openly-accessible canonical sources where possible (publisher pages, institutional repositories, PMC, open PDFs); flag paywalled-only links as ⚠️ Paywalled — find open alternative.
• Tier 2 publication. Keep inline Author (year) Title, Section. Name the framework AND its analytic move in the prose. Aggregator links (ResearchGate, academia.edu) only as supplementary alongside a canonical link.

Pass 9 — Tier register check

Confirm the prose matches the target audience tier.