Skip to content

SKILL.md

Latest commit

 

History

History
122 lines (85 loc) · 8.96 KB

SKILL.md

File metadata and controls

122 lines (85 loc) · 8.96 KB
Raw
name structured-communication
description Activate when: user says 'they didn't get it,' 'I don't know how to bring this up,' 'the message wasn't clear,' 'how do I give feedback without sounding critical,' needs to structure a funding pitch or difficult conversation, or wants to diagnose why a previous message failed to land. Do NOT activate when: the exchange is purely informational with no specific response required, or the situation calls for spontaneous casual conversation with no persuasion/coaching/praise goal. More: deciqai.com/s/structured-communication

Structured Communication

Overview

Most communication failure is a failure of sequence and mode — not information. The same facts in the wrong order land as noise. Six models cover the principal communication goals in professional life: RIDE (Risk-Interest-Difference-Effect, persuasion), FIRE (Fact-Interpret-Reaction-Effect, persuasion), GROW (Goal-Reality-Option-Will, coaching), SCRTV (Scene-Conflict-Reason-Tactics-Value, influence), ORER (Opinion-Reason-Example-Result, argumentation), FFC (Feeling-Fact-Contrast, praise).

Composes with [first-principles] (ground the factual slots), [door-in-the-face] (apply influence mechanics on top of sound structure), and [jobs-to-be-done] (identify what the listener hires communication to do, then select the matching model).

When to Use

  • Before high-stakes messages: funding pitch, difficult feedback, proposal to skeptical stakeholder, coaching session, all-hands.
  • When a previous message failed to land and you need to diagnose why.
  • When you know what to say but are unsure how to sequence it.
  • When NOT to use: purely informational exchanges with no specific response required; casual social contexts; when factual content is insufficient; when the situation requires listening; emergency/crisis communication.

Coaching Novices (Adaptive Front Door)

  • Engine mode: user has a concrete scenario → diagnose goal, select model, fill all slots, return Blueprint.
  • Coach mode: user is unfamiliar or has no concrete case → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

  1. What it is: Structured Communication picks the right model for your goal — persuade, solve together, express, or praise — and fills each slot with specific content before you speak or write.
  2. Check fit: "What outcome do you need from the other person — a decision, behavior change, understanding, a feeling of being valued? Who is the audience and what do they resist?"
  3. Elicit the real case: "(a) persuade → RIDE or FIRE; (b) coach/solve → GROW; (c) explain complex situation → SCRTV; (d) logical argument → ORER; (e) genuine praise → FFC."

[WAIT — do not advance until user responds]

  1. Fill slots: Work through each slot with the user — "What specifically goes here? Is it concrete or still abstract?"

[WAIT — do not advance until user responds]

  1. Close: "If every slot is filled with specific content and the sequence is followed, the receiver will [feel X / understand Y / decide Z]."

[WAIT — do not advance until user responds]

The Process

Stop-rule: If you cannot fill a slot with specific, concrete content — not a category label — stop. The model reveals a gap. A model with generic slots is an outline, not a Blueprint.

  1. Diagnose goal: Choose one — (a) persuade; (b) coach/solve; (c) explain and create buy-in; (d) argue logically; (e) give genuine praise.
  2. Map to model: (a)→RIDE or FIRE · (b)→GROW · (c)→SCRTV · (d)→ORER · (e)→FFC. Justify the selection.
  3. Fill every slot with specific content — content a third party could verify or act on.
  4. Check the emotional arc — does the sequence create the intended progression?
  5. Integrate resistance — address the most predictable pushback inside the structure, not deferred to "if they push back."
  6. Rehearse — note deviations; each is a potential failure point.

Model Slot Reference

RIDE: R(Risk — specific/vivid) · I(Interest — good specifically for them) · D(Difference — vs. named alternatives) · E(Effect — quantified)

FIRE: F(Fact — verifiable, not inference) · I(Interpret — own it: "What this means to me...") · R(Reaction — emotional acknowledgment) · E(Effect — call to action)

GROW: G("What would a good outcome look like?") · R("What's happening now? What have you tried?") · O("What else could you do?" — multiple before evaluating) · W("What will you do, and by when?" — commitment not intention)

SCRTV: S(Scene — context before conflict) · C(Conflict — named after Scene) · R(Reason — root cause) · T(Tactics — solution) · V(Value — outcome)

ORER: O(Opinion — conclusion first) · R(Reason — logical basis) · E(Example — specific/concrete) · R(Result — conclusion restated, closes arc)

FFC: F(Feeling — genuine reaction first) · F(Fact — specific observable behavior) · C(Contrast — vs. peer standard or alternative)

Output: Communication Blueprint

Communication goal: [Persuade/Solve/Express/Praise] | Receiver: [who, what they know/resist/care about]
Model selected: [RIDE/FIRE/GROW/SCRTV/ORER/FFC] | Justification: | Mode: [Spoken/Written]
Filled Slots: [each slot with specific content]
Resistance: Most likely: | Integrated at slot [X]:
Arc check: [intended progression confirmed]
Rehearsed: [Yes/No] | Deviations:

→ Method in Action: Lincoln's Cooper Union Address (1860)

Communication Packs

  • Investor/Fundraising: RIDE primary + ORER for written memo. Engineering/Technical: ORER (conclusion first, logic-led).
  • Management/Feedback: FIRE if emotional; FFC if goal is appreciation. Coaching: GROW — every instinct to lecture becomes a question; press W for specific commitment with deadline.
  • All-Hands/Org: SCRTV — large audiences need Scene before Conflict; context makes conflict intelligible.

Applying It Well

  1. Diagnose goal before model — defaulting to ORER when RIDE is needed is the most common error.
  2. Every slot must be specific — verifiable by a stranger reading it cold.
  3. GROW: press Will for a specific commitment with deadline, not vague intention.
  4. FFC: Feeling must be genuine — fabricated reactions are detected within seconds.
  5. FIRE: Reaction slot is the most skipped and most important — skip it and the receiver feels unseen.

→ Primary sources: references/sources.md

Common Rationalizations

[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.

Fake move Reality
[D] Using ORER when the goal is persuasion and receiver is skeptical. Skeptical receivers need RIDE or FIRE — risk/interest framing creates motivation logical argument alone cannot.
[D] Filling FIRE's Fact slot with an inference rather than observable fact. "You seemed disengaged" is inference. "You didn't respond to three emails" is a fact. The whole sequence fails if Fact is disputed.
[D] Using FFC without genuine Feeling. Generic praise without genuine emotional reaction is detected as performative and discounted.
[D] Applying GROW and then telling the receiver the answer when they take too long. The moment the coach answers for the receiver, GROW becomes ORER-as-dialogue — which produces resistance.
[D] SCRTV but leading with Conflict before Scene is established. Audience without context interprets Conflict as alarm or noise. Scene makes Conflict interpretable.
[D] "I know the structure; I don't need to write it out." Unwritten blueprints are unverifiable. Under pressure, communicators revert to default patterns.
→ Add [O] entries here after each real use — paste the actual failure pattern What went wrong and why

Red Flags

  • Cannot name the goal · Model selected without justification · FIRE's Fact is an inference not an observable.
  • GROW's Will accepted as vague intention · FFC leads with Fact not Feeling · SCRTV Conflict before Scene.
  • Blueprint not rehearsed before delivery.

Verification

  • One goal selected (not combined) · Model justified · Every slot specific and concrete.
  • Stop-rule applied: unfillable slots triggered a pause to gather missing content.
  • Resistance integrated at a specific slot · Emotional arc confirmed.
  • Blueprint rehearsed in actual delivery mode · Complete and reviewable by a third party.

Part of deciqAI Knowledge Skills — 233 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/s/structured-communication · Built by deciqAI · github.com/deciqAI · Contributions welcome.

Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/structured-communication.json