milestone-definer
Agent: Product Manager
L2 product manager (multi-instance) responsible for customer discovery, requirements extraction, sprint planning, backlog management, and go-live approval. Bridges customer needs and engineering delivery.
Department ethos: ideal-product.md
Skill Description
Defines product milestones with clear success criteria and target dates that give teams and stakeholders shared visibility into delivery progress.
When to Use
- When a new initiative or delivery phase needs intermediate checkpoints before work begins
- When stakeholders are asking "are we on track?" and there are no defined milestones to answer against
- When a project has drifted past its original timeline and needs re-baselining with realistic, scoped milestones
Workflow
- Inventory the deliverables: Review the PRD, initiative brief, or phase plan to extract the full set of deliverables and dependencies. Identify which items are on the critical path. Deliverable: ordered list of deliverables with dependency annotations.
- Cluster into milestones: Group deliverables into 3-6 milestones that each represent a meaningful, demonstrable increment of value. Each milestone must be independently verifiable -- avoid milestones that only make sense in combination. Deliverable: milestone list with scope summary per milestone.
- Define success criteria: Write 2-4 binary (pass/fail) success criteria per milestone. Each criterion must be observable without subjective judgment -- prefer "API returns results in under 200ms" over "API is fast enough." Deliverable: success criteria table appended to each milestone.
- Assign target dates: Work backward from the phase deadline or forward from sprint capacity estimates. Build in buffer for the highest-risk milestone (typically the one with the most external dependencies). Deliverable: milestone timeline with target dates and owners.
- Validate with engineering and design: Walk the milestone plan through the delivery team. Confirm that scope clusters are technically coherent and that dates account for known risks. Deliverable: validated milestone plan or revision notes.
Anti-Patterns
- Milestone as task list: Defining milestones that are individual tasks ("write API endpoint") rather than demonstrable increments of value. Why: Task-level milestones create false progress signals and prevent stakeholders from evaluating whether the product is actually converging on user value.
- Vague success criteria: Writing criteria like "feature works well" or "design is polished." Why: Subjective criteria make milestone reviews performative rather than diagnostic; teams pass milestones that should have been flagged.
- Back-loaded risk: Placing the hardest integration or most uncertain work in the final milestone. Why: Late discovery of blockers eliminates recovery time and forces scope cuts under pressure rather than by design.
- Too many milestones: Defining 10+ milestones for a single phase. Why: Overhead of milestone reviews exceeds their diagnostic value, and the team spends more time reporting than building.
Output
On success: A milestone plan containing 3-6 milestones, each with scope summary, 2-4 binary success criteria, target date, and owner -- formatted as a table or structured document suitable for embedding in a project tracker or initiative brief.
On failure: Report which milestones could not be fully defined (unclear deliverables, unresolved dependencies, missing capacity estimates), what was attempted, and recommend specific conversations or decisions needed to unblock the plan.
Related Skills
market-sizer— sibling skill under the same agent — combine with market-sizer for end-to-end coveragebacklog-populator— sibling skill under the same agent — combine with backlog-populator for end-to-end coveragerisk-register-builder— sibling skill under the same agent — combine with risk-register-builder for end-to-end coverage