L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Productivité · SKILL.md

program-status-report

program-status-report

Write a structured weekly or bi-weekly program status report for stakeholders. Use when the user says "write a status report", "program update", "stakeholder update", "weekly status", "RAG status", "program health report", "what do I send to leadership", "write the update for [program]", or needs to communicate delivery progress - even if they don't explici…

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: program-status-report description: > Write a structured weekly or bi-weekly program status report for stakeholders. Use when the user says "write a status report", "program update", "stakeholder update", "weekly status", "RAG status", "program health report", "what do I send to leadership", "write the update for [program]", or needs to communicate delivery progress - even if they don't explicitly say "status report". Produces a concise, scannable report with RAG status, blockers, and decisions needed.

Overview

A program status report has one job: give stakeholders the information they need to make decisions or remove blockers, in under 3 minutes. The failure mode is burying the lead - putting good news first and hiding the red items in paragraph 4. The format here front-loads risk and decisions needed, then provides supporting detail. Stakeholders who need detail can read further; those who only need the headline get it immediately.

Workflow

Step 1: Gather inputs

Ask the user for:

  • Program name and reporting period (e.g., "Week of March 3")
  • Overall RAG status (Red / Amber / Green) - if they are not sure, work through it in Step 2
  • Key accomplishments since last report (what shipped, what was decided)
  • Current blockers (what is slowing or stopping the program)
  • Decisions needed from stakeholders (specific asks, not general concerns)
  • Upcoming milestones (next 2-3 checkpoints and dates)
  • Any risks that surfaced or changed this period

If the user provides a narrative or notes, extract these fields from their input before generating the report.

Step 2: Determine the RAG status

If the user is unsure of the overall status, use this rubric:

Green: Program is on track. No unresolved blockers. Milestones will be met as planned.

Amber: Program is at risk. One or more of:

  • A milestone is at risk but there is a recovery plan
  • A dependency is unresolved but expected to resolve within 1 sprint
  • Scope reduction is being discussed but not yet decided

Red: Program is off track. One or more of:

  • A milestone will be missed with no recovery plan
  • A dependency is unresolved with no clear path forward
  • Scope or date change is required and has not been approved
  • A critical team or resource issue has no mitigation

Amber should be used before a problem becomes Red - it is an early warning, not a failure state. Do not use Green if anything requires stakeholder action.

Step 3: Write the report

Use this template exactly:

PROGRAM STATUS - [your program]
Period: [week / sprint / date range]
Updated: [date]
Owner: [your name]

STATUS: [GREEN / AMBER / RED]
[One sentence explaining the status. If Amber or Red, name the specific issue.]

---

DECISIONS NEEDED
[If none, write "None this period."]
1. [Decision needed from whom, by when, and what happens if it is not made]
2. [Next decision if any]

BLOCKERS
[If none, write "No active blockers."]
1. [Blocker description] - Owner: [name] - Target resolution: [date]
2. [Next blocker if any]

---

ACCOMPLISHMENTS (this period)
- [What shipped, what was accepted, what was decided]
- [Keep to 3-5 bullets max. Not every ticket - program-level progress only]

UPCOMING MILESTONES
| Milestone | Target Date | Status |
|-----------|------------|--------|
| [name] | [date] | On track / At risk / Delayed |
| [name] | [date] | [status] |

---

RISKS (new or changed)
[If none, write "No new risks this period."]
- [Risk]: [likelihood] / [impact] - Mitigation: [action and owner]

NOTES / CONTEXT
[Optional: 2-3 sentences of context that stakeholders need. Skip if nothing material.]

Step 4: Apply the three-minute rule

After writing the report, verify:

  • A stakeholder can read the STATUS line and DECISIONS NEEDED section in 30 seconds and know if action is required from them
  • The report is 1 page max when pasted into Slack, email, or a wiki
  • Accomplishments section is 3-5 bullets - not a sprint changelog
  • If anything in the Blockers or Decisions section is more than 2 lines, it is too long - link to a separate doc

Step 5: Calibrate frequency and audience

Recommend the right distribution based on program phase:

  • Weekly: standard for active programs with external dependencies or stakeholder decisions pending
  • Bi-weekly: appropriate for programs in steady-state delivery with no active escalations
  • Daily: only when a program is Red and actively in recovery mode

Audience guidance:

  • Executives / leadership: STATUS + DECISIONS NEEDED only. They do not need the milestone table.
  • Peer teams / dependencies: BLOCKERS + UPCOMING MILESTONES. They need to know what you need from them.
  • Full team: Full report including accomplishments and notes.

Anti-Patterns

1. Green-washing - everything is always green Bad: Status is Green every week until the release is 2 weeks late. Good: Use Amber early. Amber means "we have a plan but it needs attention" - that is valuable signal, not failure.

2. Accomplishments section as a ticket dump Bad: 15 bullets listing every closed Jira ticket. Good: 3-5 program-level progress statements. "Core checkout flow accepted by QA" not "JIRA-1234, JIRA-1235, JIRA-1236 closed."

3. Vague blockers Bad: "We're blocked on infrastructure." Good: "Waiting on cloud team to provision staging environment. Owner: [name]. Target: [date]. If not resolved by [date], release date is at risk."

4. Decisions buried at the bottom Bad: Status report where the decision needed from leadership is in paragraph 6. Good: DECISIONS NEEDED is the second section - immediately after the RAG status line.

5. No cadence - status reports only when things go wrong Bad: Stakeholders only hear from you when there is a crisis. Good: Regular cadence trains stakeholders to expect updates and creates a paper trail. It also normalizes Amber status before it becomes Red.

Quality Checklist

  • RAG status is on the first line with a one-sentence explanation
  • DECISIONS NEEDED section is above accomplishments and milestones
  • Every blocker has a named owner and target resolution date
  • Accomplishments are 3-5 bullets of program-level progress, not ticket lists
  • Report is skimmable in under 3 minutes
  • If status is Amber or Red, a recovery plan or escalation path is stated
Skills enregistrées