How to Explain a W3C Specification Section from Pasted Text
Spec explainers upgrade MAY to MUST and mint an errata ID. Quote the pasted RFC2119 words. Leave the rest NOT IN PASTE.
The matching generator is the W3C Specification Section Explainer from Pasted Spec Text (No Invented Errata) (W3C Specification Section Explainer From Pasted Spec Text No Invented Errata) prompt. Browse related cards in the PromptDig library (Browse more prompts). When a filled run survives, share the version you actually use (Share a prompt).
Map every claim to a short quote
Explain a pasted W3C specification section in plain language. No invented errata, MUST/SHOULD promotions, or extra algorithms. Start by filling Inputs, not by asking the model to remember last week's run. If a field is blank, write NONE or NOT IN INPUTS and leave it blank through Generate. The card is built so the model cannot honestly invent a number, owner, URL, or command that you did not paste.
Paste these fields before you hit run:
Pasted spec text: [Spec]
Section id or heading as pasted (or UNKNOWN): [Section]
Spec name and date I allow (or UNKNOWN): [Name]
Words I must not use: [Banned]
What I must never invent (errata, extra MUSTs): [Never]
Output format: [Format]
Language: [Lang]
Audience I may name: [Audience]
RFC2119 words I may keep only if present: [Normative]
What this is not: [Not]
That inventory is the honesty ledger. Anything that does not appear there is forbidden in the draft. If you catch yourself adding a nice-to-have after the run, you are no longer using the card. You are ghostwriting. Put the extra fact in Inputs and run again.
Keep MUST SHOULD MAY only if present
Generate is numbered on purpose. Do not skip a step because the first paragraph looked done. The early steps exist to stop later prose from smuggling claims.
Walk the Generate list in order:
- Banner: not errata, not a bug report, not legal advice. Quote Not.
- Honesty ledger: Name, Section, Audience, Lang, every RFC2119 word actually in Spec. Forbidden: invented errata IDs, extra MUST, other headings.
- Quote map: short quotes from Spec for each claim you will explain. If a claim has no quote, drop it.
- Plain explainer: section-by-section restatement. Keep MUST/SHOULD/MAY only if they appear in Spec.
- Out of scope: list neighboring algorithms not in Spec as NOT IN PASTE.
- Refuse: invented errata, 'browsers must already do X' if X is not in Spec, IETF RFC number invent.
- Never: do not upgrade MAY to MUST.
- Compliance pass: quote Banned and Never hits. Cut them. Format as Format.
If a step asks for a version lock, quote the version from Inputs in the output. If a step asks for a refuse list, keep the refuse list in the published artifact, not in a sidebar you delete. Reviewers should see what the model was not allowed to do.
Do not invent neighboring sections
Most failures are the same shape: a missing field gets a confident fill. A conversion rate appears. A Gradle task appears. A flash point appears. A caption appears on a job that asked for slide text only. Your review is to search the draft for numbers, names, and commands, then grep Inputs. No match means cut.
Honor the constraints as hard stops, not vibes:
- W3C section explainer from Spec. Not IETF errata and not legal advice.
- Never invent errata or promote MAY to MUST.
- Only Section named in Inputs.
- No emojis.
When the card says not legal advice, not certification, not an exam dump, or not a caption engine, that sentence belongs at the top of the output. Deleting it to look more finished is how you inherit risk.
Refuse errata IDs and IETF numbers
Finish with the compliance pass the prompt already asks for. Quote the banned-word hits. Cut them. Print character counts when the job has a cap. Print word counts when the job has a budget. List gaps as gaps. Five missing facts are more useful than one smooth paragraph.
Tags on the card (w3c specification section explainer, pasted spec no invented errata, plain language w3c section) are a reminder of the job shape, not an invitation to wander into a neighboring cluster. If you need a different surface, open a different PromptDig card rather than stretching this one.
Fill the card, then run
Replace every bracket. Run on ChatGPT, Claude, or Gemini. Read the ledger first, then the artifact. If the model invents a commit, KPI, DOI, PEL, bid, or logo, discard the run. Tighten Inputs. Run again. Share the filled card that survived, not the first draft that sounded done.