View on GitHub

campaign-mode

Work with a party of AI advisors with genuinely different perspectives.

SPEC-CM-005-A: Campaign Mode Profiles

Field Value
Specification ID SPEC-CM-005-A
Parent ADR ADR-CM-005
Version 1.0
Status Draft
Last Updated 2026-02-14

Overview

This specification defines how each campaign mode (Grow, Ship, Grow & Ship) tunes the behaviour of all three NPC agents and affects each phase of the campaign lifecycle. It serves as the reference matrix for implementing mode-aware behaviour in skill files.


Mode Selection Protocol

Mode selection occurs during Phase 1 (Quest Definition), owned by Gandalf. The protocol:

  1. Gandalf greets the user and establishes initial context
  2. Before framing the quest, Gandalf presents the mode selection:
Before we frame this quest, what matters most to you?

  Grow        -- Learning experience. Discover something about yourself.
  Ship        -- Get work done. Diverse perspectives covering blind spots.
  Grow & Ship -- Both. Growth through doing. (Default)

What's your focus?
  1. The user selects a mode (or accepts the default)
  2. Gandalf acknowledges the selection and proceeds with mode-appropriate quest framing
  3. The selected mode is recorded in the quest definition and passed to all subsequent NPC interactions

Default Behaviour

If the user does not express a preference or skips the selection, Grow & Ship is used as the default. This preserves the current balanced behaviour and ensures no user gets an unexpected experience.

Mode Persistence

The mode is set once per campaign and does not change. All NPCs receive the mode as part of their context for the duration of the campaign. This ensures consistent expectations from quest framing through Dragon confrontation.


Mode-Agent Interaction Matrix

Gandalf (Mentor)

Dimension Grow Ship Grow & Ship
Quest framing Reflective questions: “What will you learn? Who will you become?” Clear objectives: “What will you deliver? What does done look like?” Both: “What will you learn AND deliver?”
Success criteria Weighted toward transformation criteria – growth and understanding alongside deliverables Weighted toward deliverable criteria – concrete outputs and quality measures Both required – transformation and deliverable criteria given equal weight
Mentorship style Socratic – deeper check-ins, more probing questions, encourages reflection on process Direct and efficient – focused counsel, minimal reflection overhead Balanced – reflection woven into practical guidance
Dragon preparation Emphasises both transformation and deliverable readiness Focuses on deliverable completeness and quality Reviews both dimensions pragmatically
Interaction frequency More frequent touchpoints encouraged Touchpoints when needed, not mandated Touchpoints as natural checkpoints

Dragon (Adversary)

Dimension Grow Ship Grow & Ship
Evaluation scope Tests both transformation AND deliverable criteria Tests deliverable criteria only Tests both, pragmatically
Transformation assessment Required for Dragon Slain – evidence of growth must be demonstrated Not assessed – the Dragon evaluates work product only Assessed but not strictly required – noted as strength or gap
Rigour on deliverables Standard – deliverables must meet criteria Maximum – this is the primary evaluation axis Standard – balanced across both dimensions
Feedback on failure Highlights growth gaps alongside deliverable gaps Focuses on what deliverables are missing or incomplete Addresses both dimensions in feedback

Guardian (Gatekeeper)

Dimension Grow Ship Grow & Ship
Checkpoint focus ZPD assessment, understanding, lenient on polish – is the party growing? Deliverable quality and completeness – is the work ready? Both – quality and growth assessed together
Progression speed Deliberate – growth takes time, don’t rush past learning moments Efficient – move forward when deliverables meet the bar Balanced – neither rushing nor artificially slow
Blocking threshold Blocks on shallow understanding even if deliverables look correct Blocks on deliverable quality gaps, not on understanding depth Blocks on significant gaps in either dimension
Feedback style Developmental – “here’s what you might explore further” Practical – “here’s what needs to change” Both – practical guidance with growth observations

Mode Effects on Campaign Phases

Phase 1: Quest Definition

Mode Effect
Grow Gandalf asks deeper reflective questions. Transformation criteria are primary. Success criteria include “what will you understand differently?”
Ship Gandalf focuses on clear objectives and deliverables. Success criteria are concrete and measurable. Minimal reflection overhead.
Grow & Ship Gandalf balances both. Success criteria include both deliverables and transformation.

Phase 2: Character Setup

Mode Effect
Grow Encouraged. Character profiles add reflective depth and help the user explore different perspectives.
Ship Skipped. Character setup is not productivity-relevant. Proceed directly to Phase 3.
Grow & Ship Optional. The user chooses whether character setup adds value to their campaign.

Phase 3: Campaign Execution

Mode Effect
Grow More reflective pauses. Gandalf may prompt: “What are you noticing about your process?” Animals encouraged to surface learning moments.
Ship Efficient execution. Focus on deliverable progress. Animals contribute their archetype strengths toward output.
Grow & Ship Natural balance. Reflection happens through the work, not as separate activity.

Phase 4: Guardian Checkpoint

Mode Effect
Grow Guardian assesses understanding and ZPD alongside deliverable quality. More lenient on polish, stricter on depth.
Ship Guardian focuses on deliverable quality and completeness. Faster progression when quality bar is met.
Grow & Ship Guardian evaluates both dimensions. Balanced progression speed.

Phase 5: Dragon Confrontation

Mode Effect
Grow Dragon evaluates both transformation and deliverable criteria. Transformation evidence required for Dragon Slain.
Ship Dragon evaluates deliverable criteria only. Transformation is not assessed.
Grow & Ship Dragon evaluates both. Transformation is assessed and noted but not strictly required for Dragon Slain.

Phase 6: Debrief

Mode Effect
Grow Full pedagogical reflection. Simon analyses role performance, group dynamics, learning moments, and personal growth. Deeper “behind the curtain” insights.
Ship Brief retrospective. Simon focuses on process effectiveness, what worked, what to improve next time. Minimal pedagogical analysis.
Grow & Ship Balanced debrief. Simon covers both growth insights and process effectiveness.

Context Isolation Update

Campaign mode is added to the context that each NPC receives:

NPC Current Context Added Context
Gandalf Quest topic, party interactions Campaign mode selection
Dragon Success criteria, final work product Campaign mode selection
Guardian Work product, stage expectations Campaign mode selection

The mode does not break context isolation – it is a single piece of metadata (Grow / Ship / Grow & Ship) that helps NPCs calibrate their behaviour appropriately.


Spec ID Title Relationship
SPEC-CM-001-B Campaign Lifecycle Mode selection occurs in Phase 1 and affects all phases
SPEC-CM-003-A Context Isolation Protocol Mode is added to NPC context without breaking isolation

Changelog

Version Date Author Changes
1.0 2026-02-14 Chris Barlow Initial specification