Audience Translation Builder

# Audience Translation Builder

Generate `audience_translation.md` for a social-agent skill package.

No roleplay requirement:
- Do not imitate social identity or job title voice.
- Translate one idea across audience knowledge levels.

Required sections:
1. `Audience Segments`
2. `What To Emphasize`
3. `What To Avoid`
4. `Vocabulary Shift`
5. `One-Sentence Version`
6. `Manager / Decision-Maker`

Expression-focused semantics:
- Keep the claim constant and vary framing, vocabulary, and detail depth.

Minimum audience set:
- `Novice`
- `Cross-functional peer`
- `Domain expert`
- `Skeptical listener`
- `Manager / Decision-Maker`

Empty-state fallback text:
- `Audience Segments: Use default four segments until custom segments are provided.`
- `What To Emphasize:`
  - `Novice -> core meaning and concrete use case.`
  - `Cross-functional peer -> tradeoffs and collaboration impact.`
  - `Domain expert -> assumptions, boundary conditions, and precision terms.`
  - `Skeptical listener -> evidence limits and falsifiable checks.`
- `What To Avoid:`
  - `Novice -> jargon-heavy compression.`
  - `Cross-functional peer -> domain-only shorthand without context.`
  - `Domain expert -> oversimplified claims that hide assumptions.`
  - `Skeptical listener -> certainty claims without support.`
- `Vocabulary Shift: Prefer plain language; introduce jargon only with definition.`
- `One-Sentence Version: Explain the same idea in one sentence per segment using placeholder claim wording.`
- `Manager / Decision-Maker: Use plain-language business impact first, mechanism second, and avoid document-tone phrasing.`