adaptive-communication
Use when detecting ambiguous user intent, hedging language, open-ended framing, personal context before requests, or when unsure whether user wants exploration vs direct answer. Applies to all conversations.
Security Assessment
About adaptive-communication
Adaptive Communication is a lightweight conversational-guidance skill that helps an assistant tune its communication style to a user's implicit or explicit intent. It solves a subtle but common failure mode: giving a terse, transactional answer when a user is exploring or seeking acknowledgment, or over-explaining when a user just wants a direct result. The framing spans a spectrum from explicit-transactional to implicit-relational communication and treats both as valid.
The single SKILL.md provides detection signals for high-context/relational cues (hedging language, open-ended framing, personal context before a request, trailing sentences) versus low-context/transactional cues (direct imperatives, upfront format requirements, technical terminology, bounded scope). It then prescribes response adaptations—clarifying intent first, acknowledging subtext, offering scaffolding, or getting straight to the answer—along with rules for when to ask versus answer, edge cases (cultural, neurodivergent, mixed signals), anti-patterns to avoid (being patronizing, making the adaptation visible, over-structuring relational replies), and a compact quick-reference table.
Target users are prompt designers, assistant builders, and anyone configuring an AI's conversational behavior who wants a repeatable heuristic for reading intent and matching tone. It is pure behavioral guidance with no code, scripts, or external calls, so it functions entirely as a soft-skills reference and carries no operational risk.
FAQ
What does this skill actually change?
It guides how an assistant reads user intent and adapts tone and structure—clarifying first for ambiguous/relational messages and answering directly for transactional ones.
How does it decide which style to use?
It uses detection signals: hedging, open-ended framing, and personal context indicate high-context/relational needs, while direct imperatives, format requirements, and bounded scope indicate low-context/transactional needs.
Does it require any setup or tools?
No. It is a single markdown guidance document with no scripts, dependencies, or external calls.
How does it avoid feeling patronizing?
Its anti-patterns explicitly warn against making the adaptation visible, over-acknowledging emotions, and over-structuring relational replies, and against forcing styles based on demographic assumptions.
When should the assistant ask for clarification?
When multiple valid interpretations exist, when questions imply unstated needs, or when hedging and open-ended framing combine—but not when the request is obvious.
Install adaptive-communication
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
bencium/bencium-marketplace