Draft evidence content
Skill kimsanguine/signal-to-growth/skills/draft-evidence-content
Evidence-driven skills that turn customer signals into measurable growth.
npx -y skills add kimsanguine/signal-to-growth --skill draft-evidence-contentAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 12 days oldThe repository was created 12 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Draft answer-first product content with a claim ledger that links customer evidence and authoritative sources. Use when creating evidence-based articles, FAQs, landing-page copy, comparison content, 답변형 콘텐츠, or GEO-ready drafts without publishing.
SKILL.md
2.2 KB, 399 tokens by cl100k_base, as published. Nobody here has run it
Draft Evidence Content
Create a useful answer while preserving which claims are observed, reported, inferred, recommended, or unknown.
Inputs
Require:
- audience and question;
- approved evidence IDs;
- current authoritative sources for time-sensitive facts;
- product claim policy;
- public, internal, and restricted data boundaries;
- CTA and publication approval boundaries.
Workflow
- State the reader question and desired next decision.
- Build a claim ledger before drafting.
- Link each material claim to evidence IDs or external sources.
- Label inference and recommendation rather than presenting them as observed facts.
- Write the direct answer before background detail.
- Add limitations, counterevidence, and update dates where they change interpretation.
- Remove or generalize restricted customer detail.
- Create a review checklist for accuracy, privacy, brand, legal, and CTA.
- Stop at a draft. Never publish without explicit approval.
Boundaries
- Let the model propose structure, explanation, examples, and CTA wording.
- Use deterministic checks for claim IDs, evidence references, dates, links, and private-data patterns.
- Require a person to approve product, legal, commercial, and public claims.
- Do not invent testimonials, customer quotes, benchmarks, or product capabilities.
- Do not convert internal evidence into public proof without permission.
Outputs
Create:
content-brief.mddraft.mdclaim-ledger.jsonlreview-checklist.md
Read references/output-contract.md before writing them.
Stop conditions
Stop when a core claim is unsupported, restricted evidence is required for the argument, the source is stale and cannot be refreshed, or publication is requested without a final approved target.
Verification
Run:
python3 scripts/stg.py scan-privacy draft.md
Require zero unsupported material claims and keep the publish action outside this skill.