agentleFS
bookchaowalitbookchaowalit/platform-agent-skills/product-management/SKILL.md

product-management

Skill0 starsChanged 9 days agoApache-2.0View on GitHub

Own product and UX artifacts — problem framing, personas, PRD and prioritised scope, KPIs and instrumentation, launch plans, experiments, outcome reviews, roadmap, screen inventory and user flows. Use when defining what a product does and for whom, or writing or reviewing a PRD, roadmap or KPI plan. Not for AI delivery (ai-product-delivery), backlog ordering and story acceptance (product-owner), project schedule and status (project-management), detailed requirements (business-analysis), content (content-distribution) or client proposals (client-discovery-and-proposals).

16 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write the PRD (01-prd.md in the docpack or templates/positions/product-manager/product-brief-and-roadmap.md): personas, in/out scope, …

    line 16

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Product Management

Decide what to build, for whom, and how success is measured, and make the user experience explicit enough to build and test.

Workflow

  1. Establish context: owner, users, current status, live evidence (usage, feedback, UAT findings, minutes) and the system's existing docpack.
  2. Frame the problem: user pain, business outcome, constraints, non-goals.
  3. Write the PRD (01-prd.md in the docpack or templates/positions/product-manager/product-brief-and-roadmap.md): personas, in/out scope, features with MoSCoW or P1/P2/P3 priority and phase, success metrics as KPI IDs and targets linked to the KPI plan (step 6), assumptions, risks.
  4. Roadmap: phases and milestones tied to outcomes; mark unapproved dates [ต้องยืนยัน].
  5. UX: build the screen inventory from actual routes/pages (ID, name, role, status) and draw the main user flows in mermaid; note design-system components in use. Use templates/positions/ux-ui-designer/research-and-design-handoff.md for a standalone design handoff. To show the built flows to the owner, client or a new tester, record a role tour and use-case videos with flow-video.
  6. Measure: turn the PRD success metrics into templates/sdlc/kpi-and-instrumentation-plan.md (architecture-docs) — an outcome, a leading, and a guardrail KPI for every charter objective, with formula, minimum volume, baseline, dated target, and the events that feed them. Draft it in phase 06 so G5 can approve it. It is the docpack type kpi-plan (KPI-PLAN), required for Tier A/B: save it as docs/measurement/kpi-and-instrumentation-plan.md with queries in analytics/kpi/<KPI-ID>.sql, and pick each KPI's ADR-0005 layer. The lifecycle Measure stage exits when events are live, checks 1–7 pass and a dated baseline exists.
  7. Launch: write templates/sdlc/launch-gtm-plan.md at a high level in phase 05 (type, tier, window, budget) and in full in phase 10 (audience, positioning, pricing, channels with consent, enablement, support content, staged rollout, launch KPIs). G4 approves the high-level plan; G10 approves the full plan.
  8. Learn: pre-register every A/B, pricing, or messaging test, and every rollout used to measure impact, in templates/sdlc/experiment-log.md before it starts. After release, run templates/sdlc/outcome-review.md at T+30, T+60, T+90, then on the KPI plan's cadence, and route each decision to a backlog item, experiment, change request, discovery question, roadmap change, KPI definition change, or feature retirement. The Learn stage exits when the T+30 review is decided and each decision is a backlog task, filed with the dedupe key and labels in architecture-docs docs/lifecycle/OUTCOME-REVIEW-TO-BACKLOG.md. Details: references/measure-learn-loop.md.
  9. Hand requirements detail and acceptance criteria to business-analysis, backlog ordering and item acceptance to product-owner, schedule, RAID and status to project-management, architecture to solution-architecture, and test design to qa-testing.

Decision rules

  • Prioritise by user outcome and evidence, not by feature count.
  • Metrics need a baseline source or an explicit measurement plan. Define each metric once, in the KPI plan; other documents cite its KPI ID.
  • Experiment design and decision rules are fixed before the test starts. No stopping early for a win unless a sequential method was pre-registered.
  • Say a release "caused" a KPI change only with high attribution confidence (a controlled experiment whose pre-registered rule was met with no invalidating deviation); otherwise say it "coincided with" the change.
  • Pricing, contract scope, or client commitments are owner decisions; record them as open questions unless a signed record exists.
  • Screens that exist only in design are labelled planned, not built.
  • Owner decisions 2026-10-23: tasks-api is the single backlog (D2), ADR-0003 CD with a post-deploy smoke test (D3), PostHog free tier (D4; sign-up, DPA and consent banner are still owner actions).
  • For autopilot repos, new tasks go in as autopilot:proposed; only the owner approves them to autopilot:todo.

Handoff

Report the PRD version, prioritised scope, KPI plan version and instrumentation status, launch plan status, open experiments, the latest outcome review verdict and its backlog task links, the scorecard level (app_goal_scorecard.py) for Tier A/B, roadmap, screen/flow coverage, open decisions, and who must approve them.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.