| name | plan-initiative |
| version | 1.0.0 |
| description | Activate when an Engineering Manager needs to shape a rough initiative into a clear, scoped, outcome-oriented brief before execution. |
| triggers | ["plan-initiative","initiative","kickoff","scope","framing","epics","discovery","roadmap"] |
Plan Initiative
A facilitation skill for Engineering Managers to move from a rough idea to a solid initiative brief. This skill prioritizes thinking quality over document speed: clarify the problem first, then define delivery.
When to Activate
- A user asks to plan, frame, or structure an engineering initiative
- Scope is unclear and needs boundaries before execution
- The team is jumping to solutions without a validated problem statement
- Cross-team dependencies, risk, or ownership boundaries are still fuzzy
Core Concepts
- Outcome before output: Define what change should happen before deciding what to build
- Strict clarification gate: Do not draft until mandatory fields are clear enough
- Scope discipline: In-scope and out-of-scope are equally important
- Risk transparency: Make assumptions, unknowns, and dependency criticality explicit
- Single artifact focus: Produce one initiative brief, not multiple parallel documents
Detailed Guidance
Phase 1: Deep discovery (strict-stop mode)
Start with the EM's own description, then ask focused questions in small batches. Keep iterating until uncertainty is low across all mandatory fields.
Ask for:
Problem and motivation (mandatory)
- What exact problem are we solving, for whom?
- What evidence exists (signals, incidents, trends, feedback)?
- Why now? What is the forcing function?
- What happens if we do nothing?
Desired outcome (mandatory)
- What must be different in 3-6 months?
- What decision, behavior, or capability does this unlock?
- How would we recognize progress (metric or proxy if available)?
Scope and boundaries (mandatory)
- What is explicitly in scope?
- What is explicitly out of scope?
- What are the ownership boundaries and decision rights?
Delivery shape
- Is discovery needed before build?
- Can value be delivered incrementally?
- Which parts are sequential vs parallel?
Team, dependencies, and constraints
- Which teams are involved?
- Which dependencies are critical path vs non-critical?
- What constraints exist (timeline, staffing, budget, architecture, compliance)?
Risks, unknowns, and assumptions
- Top derailers (technical, product, org)
- Unknowns that must be resolved early
- Assumptions currently treated as facts
Facilitation behavior rules
- Ask a few high-value questions at a time (avoid checklist dumping)
- Challenge vague statements; ask for concrete wording
- Detect solution-jumping and pull back to the problem
- If mandatory fields remain unclear, continue discovery and do not draft
Phase 2: Draft the initiative brief
Draft only after mandatory fields are clear enough for a credible plan.
Title rules:
- Outcome-oriented and concise (roughly <= 10 words)
- Reflects what changes, not what is built
- Avoid internal jargon
Use this format:
## SUMMARY
[1-3 sentences: what this is, why now, expected change]
## PROBLEM
[Concrete problem statement with available evidence/signals]
## OUTCOME
[What changes when successful; include metric/proxy if meaningful]
## CURRENT SITUATION
[Known context, prior attempts, constraints]
## SCOPE
[Explicit in-scope boundaries]
## OUT OF SCOPE
[Explicit exclusions]
## APPROACH
[High-level approach, phases, discovery vs build logic]
## PROPOSED EPICS
Discovery:
- [Epic title] - [one-line purpose]
Build:
- [Epic title] - [one-line purpose]
## TEAM AND DEPENDENCIES
[Teams involved, dependency map, criticality]
## RISKS, UNKNOWNS, ASSUMPTIONS
[Top risks, open unknowns, key assumptions to validate]
## NEXT STEPS
[3-5 concrete actions with immediate execution value]
Phase 3: Confirm and finalize
After showing the draft, always ask for explicit selection:
- Use as-is
- Edit something
- Cancel
Do not finalize without confirmation.
Quality gate (must pass before presenting draft)
- Problem is specific and evidence-based enough
- Outcome reflects real impact, not only shipped output
- Scope and out-of-scope are explicit and coherent
- Approach fits uncertainty level (discovery where needed)
- Epics cover a realistic path from uncertainty to delivery
- Risks and assumptions are explicit and non-generic
- Next steps are concrete and immediately actionable
- Capacity/dependency realism is reflected in the plan
Examples
Good activation input:
"Help me plan an initiative to reduce incidents caused by config drift across services."
Expected behavior:
- Ask discovery questions in batches (problem evidence, forcing function, ownership boundary)
- Continue probing until mandatory fields are clear
- Draft a single initiative brief with discovery/build epics
- Ask for explicit confirmation (use/edit/cancel)
Bad behavior to avoid:
- Producing a full plan after one vague sentence
- Treating "improve reliability" as a complete problem statement
- Forcing hard metrics when only qualitative signals exist at this stage
Guidelines
- Never draft the initiative brief while any mandatory field is materially unclear
- Ask high-value questions in small batches, not full questionnaires
- Keep output to one initiative brief unless the user explicitly asks otherwise
- Require explicit
SCOPE and OUT OF SCOPE sections in every draft
- Record risks, unknowns, and assumptions separately from scope
- Treat metrics as optional: include quantitative targets when meaningful, otherwise use clear proxies
- Always close with a 3-way confirmation gate: use as-is, edit, or cancel
- Favor concrete language over abstractions and generic phrasing
Integration
- Builds on:
recursive-exploration (iterative discovery and confidence gating)
- Related:
documentation (clear structure and precision in written artifacts)
- Complementary external framing references:
Skill Metadata
- Created: 2026-04-09
- Last Updated: 2026-04-09
- Author: didacrios
- Version: 1.0.0