View on GitHub

campaign-mode

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

SPEC-CM-010-A: Phase 3 Party Engagement

Field Value
Specification ID SPEC-CM-010-A
Parent ADR ADR-CM-017
Version 1.0
Status Draft
Last Updated 2026-02-16

Overview

This specification defines how animal agents proactively engage during Phase 3 (Campaign Execution). It covers four mechanisms: a recommended first advisor at Phase 3 entry, the Next Perspective protocol for inter-animal handoffs, proactive trigger detection per archetype, and Party Assignments mapping success criteria to animal advisors.

Together these mechanisms transform the advisory council from passive tools into an active team that guides the user through the campaign.


When Gandalf transitions to Phase 3 (after quest framing and optional character setup), Gandalf analyses the quest characteristics and recommends which animal advisor to consult first.

Selection Criteria

Quest Characteristic Recommended Advisor Reason
High risk or uncertainty Cat Risk assessment should precede action
Tight timeline or complex sequencing Owl Structure and planning before execution
Unclear direction or competing priorities Bear Vision and direction to focus effort
Low motivation or daunting scope Puppy Enthusiasm and opportunity-finding to build momentum
Resource constraints or dependencies Rabbit Resource mapping before commitment
Multi-stakeholder or team alignment needed Wolf Cohesion and buy-in before divergent work

Gandalf evaluates the quest narrative, success criteria, and anticipated dragon to select the most relevant advisor. If multiple characteristics are present, Gandalf uses judgement to select the most pressing one.

Phase 3 Entry Options

Gandalf replaces the generic Phase 3 transition with a recommended first advisor. The AskUserQuestion options are:

  1. Consult {recommended advisor} first — with a one-line reason tied to the quest (e.g., “Your quest has significant unknowns — the Cat can map the risks before you begin”)
  2. Begin working — with trigger phrases planted for mid-campaign transitions (“When you’re ready for a checkpoint, say ‘I’m ready for a checkpoint’”)
  3. Review quest summary — show the quest definition, success criteria, and campaign mode
  4. Consult a different advisor — the user wants a different animal perspective first

If the user selects “Consult a different advisor”, present the full animal selection menu (with profile names if applicable).


2. Next Perspective Protocol

After every Phase 3 animal consultation, the animal uses AskUserQuestion to suggest the next perspective. This creates continuity between consultations rather than leaving the user alone after each conversation.

AskUserQuestion Options

At the end of every Phase 3 consultation, the animal presents:

  1. Consult {suggested next animal} — with a one-line reason based on the conversation (e.g., “We’ve identified three risks — the Owl can help structure a plan around them”)
  2. Consult a different advisor — the user wants a perspective not suggested
  3. Continue working — the user is ready to work on their own
  4. Request a checkpoint — the user wants a Guardian evaluation
  5. Consult Gandalf — the user wants strategic counsel from the mentor

Archetype Complement Fallback Table

Animals adapt their suggestion to the conversation content. When conversation context does not clearly indicate a next perspective, use this fallback table:

Animal Default Suggestion Reason
Bear Cat or Owl Direction set — assess risks or structure the path
Cat Owl or Rabbit Risks mapped — plan around them or check resources
Owl Bear or Rabbit Structure ready — validate direction or identify needs
Puppy Cat or Wolf Opportunities found — stress-test or check alignment
Rabbit Owl or Wolf Resources mapped — schedule work or ensure buy-in
Wolf Bear or Puppy Alignment checked — revisit direction or build momentum

The table is a fallback, not a script. Animals should adapt based on what was actually discussed. Use profile names when profiles are assigned.


3. Proactive Engagement Triggers

Each animal archetype has trigger signals — topics or patterns in conversation that indicate another archetype’s perspective is needed. Animals use these triggers to make informed Next Perspective suggestions and to proactively address cross-archetype concerns.

Per-Archetype Trigger Signals

Archetype Trigger Signals (detect in conversation)
Bear Direction unclear, competing priorities, “what should we focus on?”, vision drift
Cat Risks mentioned, unknowns identified, “what could go wrong?”, scope concerns, assumptions surfaced
Owl Timeline discussed, sequencing questions, “how long?”, process gaps, scheduling concerns
Puppy Low energy, discouragement, missed opportunities, “this is hard”, scope for enthusiasm
Rabbit Resource constraints, dependency questions, “who do we need?”, tool/skill gaps, budget concerns
Wolf Team misalignment, stakeholder friction, “they don’t agree”, collaboration gaps, buy-in issues

Trigger Response Protocol

  1. If you detect YOUR archetype’s trigger — address it directly as part of your response. This is your core strength.
  2. If you detect ANOTHER animal’s trigger — mention the observation briefly in your response (“I notice there are resource questions here that the Rabbit could help with”) and prioritise that animal in your Next Perspective suggestion.
  3. Multiple triggers detected — prioritise the most urgent or foundational trigger in your Next Perspective suggestion. Use judgement.

4. Party Assignments

Party Assignments map each success criterion from the quest to a primary and secondary animal advisor. This creates a structured relationship between the quest’s goals and the advisory council’s strengths.

quest.md Format

Gandalf writes the Party Assignments table to .campaign/quest.md during Phase 1 (Quest Definition), between the Anticipated Dragon and Progress Log sections:

## Party Assignments

| Criterion | Primary Advisor | Secondary Advisor |
|-----------|----------------|-------------------|
| {criterion 1} | {animal} ({archetype reason}) | {animal} ({archetype reason}) |
| {criterion 2} | {animal} ({archetype reason}) | {animal} ({archetype reason}) |

Assignment Principles

Archetype-Criterion Mapping Guidance

Criterion Type Likely Primary Likely Secondary
Deliverable quality / output Owl Cat
Risk mitigation / safety Cat Owl
Vision / direction / strategy Bear Wolf
Team alignment / stakeholder buy-in Wolf Bear
Resource acquisition / dependencies Rabbit Owl
Motivation / engagement / momentum Puppy Wolf
Transformation / learning (Grow mode) Bear Puppy

This is guidance, not a fixed mapping. Gandalf adapts based on the specific quest.

How Animals Use Party Assignments

Animals read the Party Assignments table from quest.md (as part of their Campaign Awareness step). When progress is made on an assigned criterion, the animal suggests the primary advisor for the next unaddressed criterion in their Next Perspective options.


5. Context Window Impact

The animal campaign extension file grows from ~40 lines to ~100-120 lines with the addition of three new sections (Next Perspective, Proactive Engagement, Criterion Progress Awareness).

Updated Size Estimates

Content Previous Size New Size Delta
Animal Campaign Extensions ~40 lines ~100-120 lines +60-80 lines
Command Previous Total New Total Delta
start-quest.md ~480 lines ~540-560 lines +60-80
council.md ~480 lines ~540-560 lines +60-80
continue-quest.md ~950 lines ~1010-1030 lines +60-80

Worst case (continue-quest.md): ~1030 lines. The increase is proportionate to the capability gained — three new mid-phase engagement mechanisms for ~80 additional lines.


Spec ID Title Relationship
SPEC-CM-008-A Animal Campaign Extensions Base extension file that this spec extends
SPEC-CM-001-B Campaign Lifecycle Phase 3 definition and transition protocol
SPEC-CM-002-A Gandalf Agent Gandalf’s Phase 1 and Phase 3 transition behaviour
SPEC-CM-005-A Campaign Mode Profiles Profile name usage in Next Perspective suggestions

Changelog

Version Date Author Changes
1.0 2026-02-16 Chris Barlow Initial specification