magnus919_agent-skills/product-operations-and-governance/SKILL.md

name, description, license, compatibility, metadata

name	description	license	compatibility	metadata
product-operations-and-governance	Define and run product governance — recurring decision rights, intake, portfolio cadences, evidence standards, and cross-functional operating contracts. Covers six review cadences (intake, portfolio, roadmap, experiment, launch, lifecycle) with named accountable owners, minimum evidence standards per decision type, and escalation paths. Supports lightweight and high-assurance operating modes with configurable governance patterns. Use when designing a product governance model, resolving contested decisions, establishing evidence standards, recording exceptions and escalations, or building cross-functional operating contracts. Do NOT use for executive governance (capital allocation, org structure — route to chief-of-staff-methodology or strategy-frameworks), for technical delivery gates (CI/CD, release approval — route to release-engineering or spec-driven-development), or to impose a universal org chart.	MIT	Agent-agnostic — works with any agent framework supporting the Agent Skills format. No external services, proprietary tools, or runtime dependencies required.	tags product-operations, product-governance, decision-rights, operating-model, review-cadence, evidence-standards, escalation, exceptions, intake, portfolio-review, product-council

Product Operations and Governance

Define and operate the recurring product governance system: who decides what, with what evidence, on what cadence, and what happens when decisions are contested or evidence is missing. This skill owns the product-level operating model — the connective tissue between product strategy, portfolio choices, experimentation, adoption, and lifecycle learning. It does not own executive governance or technical delivery gates.

Governance Boundary (Read First)

This skill owns product governance: the recurring system for intake, portfolio review, roadmap decisions, experiment review, launch decisions, and lifecycle/health review. Product governance answers: what are we building, in what order, with what evidence, reviewed by whom, on what cadence?

This skill explicitly does not own:

• Executive governance — capital allocation, org structure, strategic bets at the company level, M&A evaluation. Route to chief-of-staff-methodology for decision-memo and executive-office methods, and strategy-frameworks for strategic planning frameworks.
• Technical delivery gates — CI/CD pipelines, release approval workflows, deployment checklists, infrastructure change review. Route to release-engineering for release mechanics and spec-driven-development for specification-phase gates.

The three governance layers — product, executive, and delivery — are distinct. A product governance decision ("approve this experiment to proceed to launch review") is not an executive decision ("allocate $2M to the payments platform") and not a delivery gate ("the deployment pipeline must pass integration tests"). See references/discovery-brief.md for the full boundary analysis.

Core Framework

Two Operating Modes

This skill supports two modes; choose one explicitly for every engagement. The mode determines evidence requirements, review formality, and escalation thresholds.

Dimension	Lightweight	High-Assurance
Team size	Small (≤15 engineers, ≤3 product teams)	Any size, with regulatory or safety obligations
Review formality	Async written updates; synchronous only for contested decisions	Synchronous reviews with documented quorum
Evidence minimum	Hypothesis + qualitative signal or single quantitative metric	Statistical evidence, risk analysis, compliance sign-off
Exception tracking	Team wiki or decision log	Formal exception register with revisit dates
Escalation path	Direct to accountable executive	Formal escalation chain with documented resolution
Cadence	Bi-weekly or monthly	Weekly or per-release-cycle
Artifact retention	Lightweight (spreadsheet, shared doc)	Auditable (versioned records, immutable log)

The mode is a configuration choice, not a maturity level. A startup building a non-regulated consumer app operates in lightweight mode. A medical-device team of 8 operates in high-assurance mode. A 200-person platform team may operate parts of its portfolio in lightweight mode and parts in high-assurance.

Configurable Governance Patterns

No single org chart or governance model is imposed. The skill provides configurable patterns; select and adapt:

Pattern	When to use	Key trait
Single accountable owner	Small team, single product	One person decides; reviews are advisory
Product council	Multi-team, multi-product	Cross-functional group with defined voting/consensus rules
Tiered review	Portfolio with varied risk	Lightweight for low-risk; high-assurance for regulated
Delegated authority with escalation	Scaled organization	Decision rights pre-delegated by category; escalate only exceptions

Every pattern requires the same outputs: a decision-rights map, review cadences with evidence standards, and escalation paths. Use templates/operating-model.md to capture the selected pattern and configuration.

Decision Rights

A decision-rights map answers five questions for every decision type:

1. Who decides? Named role (not "engineering" or "leadership" — a specific accountable owner).
2. Who must be consulted? Roles or individuals whose input is required before the decision.
3. Who must be informed? Roles or individuals who are notified after the decision.
4. What evidence is required? The minimum evidence standard for this decision type (varies by mode).
5. What is the escalation path? Who resolves it when the accountable owner cannot decide or the decision is contested.

Decision types the skill covers:

Decision type	Typical cadence	Lightweight evidence	High-assurance evidence
Intake accept/reject	Per-request (continuous)	Problem statement + one signal	Problem statement, cost of delay, strategic alignment score, capacity check
Portfolio prioritization	Monthly or quarterly	Relative rank with rationale	Ranked with cost-of-delay, strategic alignment, capacity model, risk assessment
Roadmap commitment	Per-planning cycle	Hypothesis + success criteria	Hypothesis, experiment results or market evidence, dependency map, confidence interval
Experiment proceed/stop	Per-experiment	Guardrail check + qualitative signal	Statistical analysis, guardrail verification, ethics review, decision rule
Launch go/no-go	Per-launch	Readiness checklist + stakeholder sign-off	Full readiness evidence packet, risk acceptance sign-off, rollback plan verified
Lifecycle continue/invest/harvest/retire	Per-review cycle	Usage + outcome data, team recommendation	Usage, financial, competitive, and risk data; multi-stakeholder review

Use templates/decision-rights-map.md to document the map for a specific product or portfolio.

Review Cadences

Six recurring reviews form the product governance rhythm. Each review has a defined purpose, participants, inputs, outputs, and decision authority.

Review	Purpose	Typical participants	Key inputs	Key outputs	Decision authority
Intake / Opportunity review	Decide which new work enters the product system	Product lead, engineering lead, design lead (varies by pattern)	Problem statement, strategic alignment, rough sizing	Accept/reject/defer decision, assigned owner	Product lead (or council vote)
Portfolio review	Sequence and resource-allocation across the portfolio	Product council or leadership group	Bet records, capacity model, strategic priorities	Prioritized portfolio, resource allocations, deferrals	Product council or accountable exec
Roadmap review	Commit, adjust, or defer roadmap items; review evidence updates	Product lead, engineering lead, key stakeholders	Updated bet records, new evidence, dependency status	Updated Now/Next/Later, continue/pause/kill decisions	Product lead with stakeholder input
Experiment review	Decide whether experiment results support proceeding, iterating, or stopping	Product lead, data/science lead, engineering lead	Experiment readout, guardrail report, decision recommendation	Proceed/stop/pivot decision, updated bet record	Product lead (with science input)
Launch review	Confirm readiness to ship; accept residual risk	Product lead, engineering lead, QA, security, support, marketing	Readiness evidence packet, risk register, rollback plan	Go/no-go/defer decision, accepted risks	Product lead (go/no-go); risk acceptance may require exec
Lifecycle / Health review	Assess product health; decide continue/invest/harvest/retire	Product lead, engineering lead, support, finance (high-assurance)	Usage data, outcome metrics, cost data, competitive intel	Lifecycle decision, updated investment level, migration plan if retiring	Product council or accountable exec

Use templates/review-cadence.md to configure cadences for a specific operating model. Routes to product-roadmapping-and-portfolio for roadmap review mechanics, product-experimentation for experiment review methods, and product-lifecycle-learning (prose — same-wave skill, directory not yet created) for lifecycle/health review evidence.

Evidence Standards

Every decision type has a minimum evidence standard. The standard scales with the operating mode. Evidence is classified into four categories in every artifact:

• Observed — measured, verified, reproducible data.
• Inferred — conclusion from observed data with stated assumptions and confidence.
• Asserted — stakeholder claim not yet verified; treated as an assumption.
• Committed — a decision with consequences for reversal; recorded with accountable owner and revisit trigger.

Missing required evidence is not a reason to skip a review — it is a reason to escalate. A review that proceeds without required evidence must produce an exception record, not silent approval.

Exceptions and Escalations

Exception Record

When a governance requirement is waived or deferred, record the exception. Without a record, the exception becomes the new default.

Fields: what was excepted, why, who approved, date approved, when to revisit (specific date or trigger condition), and what evidence (if any) substitutes for the waived requirement.

Use templates/exception-record.md.

Escalation Record

When a decision cannot be resolved at its designated level — because evidence is missing, stakeholders are deadlocked, or the accountable owner cannot decide — escalate. Escalation is not failure; it is the governance system working as designed.

Fields: what was escalated, to whom, why the lower level could not resolve, resolution, date resolved, and closure evidence.

Use templates/escalation-record.md.

Loading Guide

Load only the file relevant to the current task. Do not load everything at once.

File	Load when
references/discovery-brief.md	You need to understand governance boundaries, ownership, and routing across skills
templates/operating-model.md	Designing or configuring a product operating model from scratch
templates/decision-rights-map.md	Mapping decision rights for a product or portfolio
templates/review-cadence.md	Configuring review cadences with purposes, participants, inputs, outputs
templates/exception-record.md	Recording a waived or deferred governance requirement
templates/escalation-record.md	Recording an escalation through the governance system

Working Method

1. Select the operating mode

Start every engagement by choosing lightweight or high-assurance mode. Do not default to one. Ask: is this product regulated, safety-critical, or subject to external compliance obligations? If yes, high-assurance. If the team is small and the product is non-regulated, lightweight.

2. Choose the governance pattern

Select from single accountable owner, product council, tiered review, or delegated authority with escalation. Adapt, don't copy. Document the choice in the operating model template.

3. Map decision rights

For every decision type in scope, fill the decision-rights map: who decides, who is consulted, who is informed, what evidence is required, and where to escalate. Use templates/decision-rights-map.md.

4. Configure review cadences

Set the purpose, participants, inputs, outputs, and decision authority for each review. Adjust frequency to match the operating mode. Use templates/review-cadence.md.

5. Establish evidence standards

Define the minimum evidence standard per decision type, scaled to the operating mode. Record the standard in the decision-rights map. Evidence standards are not aspirational — they gate the decision.

6. Record exceptions and escalations

Every exception and escalation gets a dated record with accountable owner and revisit trigger. Exception records prevent waiver-by-neglect. Escalation records make the governance system observable and improvable.

Routing Table

When you need...	Load this skill
Product vision, North Star, competitive positioning	product-strategy
Tactical prioritization (RICE, MoSCoW), decision logs, specs	product-methodology
Outcome roadmaps, strategic bets, portfolio sequencing	product-roadmapping-and-portfolio
Experiment design, method selection, guardrails, readouts	product-experimentation
Post-launch learning, lifecycle decisions, assumption updates	product-lifecycle-learning (prose — same-wave skill)
Executive decision memos, CoS methods, board materials	chief-of-staff-methodology
Strategic planning, capital allocation, OKR frameworks	strategy-frameworks
Release mechanics, deployment pipelines, rollback plans	release-engineering
Specification-phase gates, acceptance criteria, task planning	spec-driven-development

When Not to Use

Do not load this skill for:

• Executive governance. Capital allocation, org structure decisions, strategic bets at the company level, or M&A evaluation — route to chief-of-staff-methodology or strategy-frameworks.
• Technical delivery gates. CI/CD pipelines, release approval workflows, deployment checklists, or infrastructure change review — route to release-engineering or spec-driven-development.
• Imposing a universal org chart. The governance patterns are configurable templates, not a mandated structure. If the ask is to design an org chart from scratch, this skill is the wrong tool.
• Single decisions without a recurring system. If you need to make one decision (not design the system for making decisions over time), use product-methodology for decision logs or adr-authoring for architecture decisions.