| name | policy-story-explainer |
| description | Rewrites long policy, compliance, HR, legal, operational, or governance documents into accurate, digestible story-led explanations. Use when the user wants a policy document made more relatable, scenario-based, easier to understand, or turned into realistic examples without watering down the original content. |
| version | 1.1.0 |
Policy Story Explainer
Purpose
Turn dry policy documents into clear, relatable explanations using realistic scenarios, plain-English summaries, and direct links back to the original policy. Improve understanding without weakening, changing, or oversimplifying the policy.
Core principles
- Preserve the original meaning, obligations, restrictions, thresholds, exceptions, definitions, and consequences.
- Do not invent policy rules that are not present in the source.
- Do not remove nuance just to make the content easier to read.
- Use realistic scenarios to explain how the policy applies in everyday situations.
- For each theme, explain why the rule exists in practical human terms: what risk, harm, failure, or obligation the policy is trying to prevent.
- Make clear what is policy, what is explanation, and what is example.
- Flag ambiguity instead of guessing.
- Avoid giving legal advice unless the source explicitly provides it.
Workflow
- Before rewriting, read references/output-format.md,
references/writing-style.md, and
references/examples.md. For long or high-stakes policies,
follow the miniature complete output in examples §6 as the shape to match.
- Read the policy carefully.
- Identify the key rules, duties, permissions, exceptions, risks, and consequences.
- Group related sections into understandable themes.
- For each theme, write: a relatable scenario, a plain-English explanation, the relevant policy rule, what the person should do, why the rule exists, and what could go wrong if ignored.
- Preserve source references such as section numbers, headings, page numbers, or clause names where available.
- Highlight any areas that need human/legal/compliance review.
Accuracy rules
These are the heart of the skill. When rewriting policy content:
- Keep "must" as "must", "shall" as "shall", "may" as "may", "should" as "should".
shall is mandatory — never soften it to "should", "will try to", or "aims to".
- Distinguish aspirational wording ("strive to", "aim to") from binding obligations ("shall", "must") even when they sit in the same clause.
- Keep numbers, dates, time limits, approval levels, and thresholds exactly as written.
- Keep defined terms and quoted definitions exactly if they affect interpretation.
- Preserve template placeholders verbatim (e.g.
[organization name], [bought, built, used, or sold]). Do not invent a name, reorder, re-inflect, or pick one option from a placeholder list.
- Preserve source references, including section numbers, clause letters, and footnote citations.
- Do not merge separate rules if doing so changes the meaning.
- Do not turn exceptions into general permission.
- Do not make a policy sound optional if it is mandatory.
- If the source is unclear, write: "The policy is not explicit on this point."
Output
For long policies, lead with a one-page executive summary (obligations, deadlines, and exceptions as scannable bullets, mandatory language intact), then the full story-led version. For a short policy, the full version alone is fine. The summary restates rules in plain English; it never drops a must/shall, a threshold, or an exception.
See references/output-format.md for the full-version template and section structure.
Writing style
Warm, clear, and human, but precise. Pitch scenarios and depth to the audience (employees, managers, or auditors), reuse one small recurring cast across a long document, and strip AI tells from the narrative sections only — never from quoted policy, definitions, or the executive summary.
See references/writing-style.md for the prefer/avoid list, the audience-tuning table, narrative-continuity rules, and the anti-slop checklist.
Examples
See references/examples.md: a single theme end to end (§0), WRONG-vs-CORRECT fidelity traps (§1), audience-tuned variants (§2), the two-tier output (§3), narrative continuity (§4), template/placeholder handling (§5), a complete miniature policy (§6), and aspire-vs-shall in one clause (§7).
When to ask for clarification
Ask the user for clarification if:
- the audience is unclear
- the document has no headings or structure
- the user wants a specific tone
- the policy has legal, safeguarding, medical, financial, or regulatory implications
- the user asks to make the policy "less strict" or "friendlier" in a way that could change its meaning
For sensitive legal, safeguarding, medical, financial, or regulatory content, surface review needs rather than resolving them yourself.
Final review checklist
Before responding, check: