technical-program-manager

어떤 스킬인가요?

Guides technical program management—multi-team initiatives with dependencies, milestones, RAID tracking, launch readiness, stakeholder status, and cross-functional coordination across engineering, product, and infrastructure (not application code or BRDs). Use when running a technical program, dependency maps, milestones, exec status, or unblocking cross-team delivery—not for requirements (business-analyst), rollout (deployment-strategist), CI/CD (devops), data roadmaps (data-manager), or single-team delivery (fullstack-software-engineer). Incidents: incident-management-engineer. Architecture: senior-system-architecture. Strategy: business-consultant. Comms: communication-lead. DC site build: data-center-design-execution-lead. DC portfolio: data-center-portfolio-planning-execution-lead. M&A/financing deal execution and closing matrix: transaction-manager. Exec/VIP and community customer escalations: community-executive-escalations-program-manager. CVD/disclosure: technical-program-manager-security-cvd.

설치 방법

프로젝트 폴더에서 아래 명령어를 실행하고, 설치할 에이전트를 선택하세요.

설치 명령을 직접 실행해야 적용됩니다. 지원 에이전트와 필요한 권한·라이선스는 제작자의 안내를 확인하세요.

스킬 원문 파일

# Dependencies and RAID

## Table of contents

1. [Dependency types](#dependency-types)
2. [RAID log](#raid-log)

## Dependency types

| Type | Example |
|---|---|
| Technical | Service B API must ship before UI integration |
| Team | Security review blocks prod deploy |
| External | Vendor API GA date |
| Environment | Staging parity with prod for load test |

Document: predecessor, successor, lead time, fallback if slip.

## RAID log

| Column | Use |
|---|---|
| **R**isk | Potential problem; likelihood/impact; mitigation |
| **A**ction | Task to reduce risk or advance program |
| **I**ssue | Active blocker; owner; target resolution |
| **D**ecision | What was decided; who; when |

Review weekly; close items with evidence, not optimism.

# Launch readiness

## Table of contents

1. [Go/no-go checklist](#gono-go-checklist)
2. [Hypercare](#hypercare)

## Go/no-go checklist

**Engineering**

- [ ] Critical path complete or waived
- [ ] Integration and e2e tests pass in staging
- [ ] Rollback tested or time-boxed recovery documented
- [ ] Feature flags or kill switches where applicable

**Operations**

- [ ] Runbooks and on-call rotation
- [ ] Dashboards and alerts for new paths
- [ ] Capacity and rate limits reviewed

**Product / GTM**

- [ ] Docs and release notes
- [ ] Support macros / training if customer-facing

**Governance**

- [ ] Security/compliance sign-off if required
- [ ] Named approvers for go/no-go recorded

## Hypercare

- 48–72h elevated monitoring after launch
- Daily stand-down during hypercare
- Retro within 2 weeks: what slipped, what to fix in process

# Program charter

## Table of contents

1. [Charter sections](#charter-sections)
2. [Phase plan](#phase-plan)

## Charter sections

| Section | Content |
|---|---|
| Problem | Why now; cost of delay |
| Goals | 2–4 measurable outcomes |
| Scope | Teams, systems; explicit out-of-scope |
| Timeline | Start, phase gates, target launch |
| DRIs | Workstream → single owner |
| Dependencies | External vendors, other programs |
| Success metrics | How success is measured at 30/90 days |

## Phase plan

Typical phases:

1. **Discover** — align scope, dependencies, estimates
2. **Build** — parallel workstreams with integration points
3. **Harden** — perf, security review, runbooks
4. **Launch** — cutover; hypercare window
5. **Steady** — handoff to BAU owner; retrospective

Each phase ends with a gate: criteria met or explicit waiver.

# Stakeholder communications

## Table of contents

1. [Audience matrix](#audience-matrix)
2. [Comms rules](#comms-rules)

## Audience matrix

| Audience | Content | Cadence |
|---|---|---|
| Workstream DRIs | Blockers, dates, dependencies | Weekly |
| Engineering leads | Technical integration, quality gates | Biweekly |
| Sponsors / exec | RAG, milestones, decisions | Biweekly or monthly |
| Support / GTM | Customer impact, timelines | Pre-launch |
| Company-wide | Only via approved channels | Launch |

## Comms rules

- Same facts everywhere; tailor depth not truth
- Bad news early with mitigation and new date
- No surprise launches—support and docs in the loop
- Record decisions in RAID, not only in chat

# Status reporting

## Table of contents

1. [Weekly status](#weekly-status)
2. [Steering summary](#steering-summary)

## Weekly status

**Header:** Program name, report date, overall RAG (G/Y/R)

**Per workstream (3–5 bullets max):**

- RAG + one-line health
- Completed since last report
- Next 7 days
- Blockers / asks (tag owner)

**Program-level:**

- Milestone tracker (planned vs actual)
- Top 3 risks
- Decisions needed from sponsors

## Steering summary

One page max:

- Objective reminder (1 sentence)
- Overall RAG + trend (↑↓→)
- Schedule: on track / slip (days) / recovered
- Budget or headcount if relevant
- Asks: decisions, staffing, scope tradeoffs

---
name: technical-program-manager
description: |
  Guides technical program management—multi-team initiatives with dependencies, milestones, RAID
  tracking, launch readiness, stakeholder status, and cross-functional coordination across
  engineering, product, and infrastructure (not application code or BRDs).
  Use when running a technical program, dependency maps, milestones, exec status, or unblocking
  cross-team delivery—not for requirements (business-analyst), rollout (deployment-strategist),
  CI/CD (devops), data roadmaps (data-manager), or single-team delivery (fullstack-software-engineer).
  Incidents: incident-management-engineer. Architecture: senior-system-architecture. Strategy:
  business-consultant. Comms: communication-lead. DC site build: data-center-design-execution-lead.
  DC portfolio: data-center-portfolio-planning-execution-lead. M&A/financing deal execution and
  closing matrix: transaction-manager. Exec/VIP and community customer escalations:
  community-executive-escalations-program-manager. CVD/disclosure:
  technical-program-manager-security-cvd.
---

# Technical Program Manager

## When to Use

- Coordinate a multi-team technical initiative with shared milestones
- Map dependencies, critical path, and integration points
- Maintain RAID log (risks, actions, issues, decisions)
- Run launch readiness or go/no-go reviews
- Produce weekly program status for leadership or steering committee
- Resolve cross-team blockers without owning the technical design

## When NOT to Use

- Write BRDs, user stories, or process maps → `business-analyst`
- Choose canary vs blue-green or write cutover runbooks → `deployment-strategist`
- Implement pipelines, GitOps, or on-call → `devops`
- Own data platform roadmap or data governance → `data-manager`
- Author service RFCs or code → `senior-software-engineer`, `fullstack-software-engineer`
- Platform product roadmap for IDP → `platform-engineer`
- Architecture decisions, ADRs, design review → `senior-system-architecture`
- Strategy, business case, operating model → `business-consultant`
- Technical discovery, solution architecture, RFP response authoring → `solutions-architect`

## Related skills

| Need | Skill |
|---|---|
| Release strategy and rollback design | `deployment-strategist` |
| Pipeline and environment operations | `devops` |
| Requirements and business cases | `business-analyst` |
| UX scope and design handoff | `product-designer` |
| Engineering design and RFCs | `senior-software-engineer` |
| Data program roadmaps | `data-manager` |
| Security/compliance program gates | `compliance-engineer`, `cybersecurity` |
| System architecture and ADRs | `senior-system-architecture` |
| Strategy and business case | `business-consultant` |
| Incident ops: SEV, on-call, postmortems | `incident-management-engineer` |
| Board resolutions, entity, corporate closing | `corporate-counsel` |
| Launch and stakeholder messaging | `communication-lead` |
| Developer training program rollout | `developer-education-lead` |
| Exec/VIP and community customer escalations | `community-executive-escalations-program-manager` |
| Data center design and facility execution | `data-center-design-execution-lead` |
| Multi-site DC portfolio planning and steering | `data-center-portfolio-planning-execution-lead` |
| M&A diligence, signing, closing coordination | `transaction-manager` |
| Coordinated vulnerability disclosure program | `technical-program-manager-security-cvd` |
| Customer solution design and delivery handoff package | `solutions-architect` |

## Core Workflows

### 1. Program charter

Define before execution:

- Objective and measurable outcomes (not output lists)
- Scope: teams, systems, in/out
- Timeline: phases, hard dates, external dependencies
- Roles: DRI per workstream, escalation path
- Success metrics and definition of done

**See `references/program_charter.md` for charter template.**

### 2. Work breakdown and dependencies

1. Decompose into workstreams (each with one DRI)
2. List deliverables per workstream with dates
3. Build dependency graph (finish-to-start; flag circular deps)
4. Identify critical path and buffer on that path only
5. Mark integration milestones (API freeze, env parity, dress rehearsal)

**See `references/dependency_risk.md` for dependency types and RAID format.**

### 3. Operating rhythm

| Cadence | Audience | Purpose |
|---|---|---|
| Weekly | Workstream DRIs | Blockers, date slips, dependency asks |
| Biweekly | Sponsors / leads | Scope, risk, decision needs |
| Pre-launch | Launch council | Readiness checklist |

- One source of truth for status (doc or tracker—not duplicate spreadsheets)
- Decisions logged with owner and date; no re-litigating closed calls

**See `references/status_reporting.md` for status template.**

### 4. Risk and issue management

- **Risk**: may happen → mitigation + trigger
- **Issue**: happened → owner + ETA + comms plan
- Escalate when: critical path slip > agreed threshold, scope creep without sponsor, unresolved cross-team conflict

**See `references/dependency_risk.md` for RAID columns.**

### 5. Stakeholder communication

- Status: RAG per workstream, changes since last week, asks
- No technical deep dives in exec updates—link appendix
- Align messaging before external comms (customers, support, legal)

**See `references/stakeholder_comms.md` for audience-specific formats.**

### 6. Launch readiness

Gate checklist before production:

- [ ] All critical-path items complete or explicitly waived by sponsor
- [ ] Rollback/runbook owned (`deployment-strategist` + `devops`)
- [ ] Monitoring and on-call briefed
- [ ] Support/docs updated
- [ ] Go/no-go with named approvers

**See `references/launch_readiness.md` for full checklist.**

## When to load references

- **Charter and phases** → `references/program_charter.md`
- **Dependencies and RAID** → `references/dependency_risk.md`
- **Status and steering** → `references/status_reporting.md`
- **Exec and sponsor comms** → `references/stakeholder_comms.md`
- **Go/no-go** → `references/launch_readiness.md`