discovery-high-level-plan

name	discovery-high-level-plan
description	Transform vague or high-level human intent into a governed Discovery action plan. Activate when intent is unclear, broad, or when multiple discovery steps are required before any artefact is created.
license	MIT
compatibility	Works with any filesystem-based AI coding agent
metadata	{"author":"gaai-framework","version":"1.0","category":"discovery","track":"discovery","id":"SKILL-DISCOVERY-HIGH-LEVEL-PLAN-001","updated_at":"2026-02-26T00:00:00.000Z","status":"stable"}
inputs	["human_intent","contexts/artefacts/** (optional)","contexts/memory/** (selective)"]
outputs	["discovery_action_plan"]

Discovery High-Level Planning

Purpose / When to Activate

Activate this skill when:

• Intent is unclear or broad
• Multiple discovery steps are required
• You don't know which skills to run first
• Refinement is needed before committing to artefacts

Typical triggers: "I have an idea for a feature", "Users are dropping during onboarding", "We should rethink pricing", "There's a bug where X happens but we don't know why"

This skill plans — it does not produce artefacts.

---

Process

1. Understand the real problem or opportunity
2. Determine Discovery scope: new initiative / feature addition / iteration / removal / clarification / bug triage (complex bug with unclear root cause)
3. Identify required artefacts
4. Select necessary skills in logical order
5. Insert validation and refinement gates
6. Output a minimal, governed plan

---

Outputs

Discovery Action Plan

Goal:
<what clarity must be achieved>

Steps:
1. <skill name> — purpose
2. <skill name> — purpose
3. <skill name> — purpose

Validation Gates:
- artefact completeness
- acceptance criteria present
- uncertainties flagged

Ready For Delivery:
yes / no

Blocking Reasons (if any):
- ...

---

Quality Checks

• Plan is minimal and focused
• All necessary artefacts will be produced
• Skill sequence is logical
• Governance is respected
• Human can approve or adjust easily

---

Non-Goals

This skill must NOT:

• Generate epics, stories, or PRDs
• Make product decisions
• Execute other skills
• Write artefact content

If the next step isn't obvious — plan first. Planning prevents waste more than speed increases output.