| name | craft-escalation-memo |
| version | 1.0.0 |
| description | Generates structured escalation documents for leadership. Evidence-based, professional, urgent without panic. |
| category | risk-management |
| trigger | Critical risk requires leadership attention, decision needed beyond team scope, blocker affecting delivery timeline. |
| autonomy | human-in-the-loop |
| portability | universal |
| complexity | advanced |
| type | generation |
| inputs | [{"name":"escalation_context","type":"structured-text","required":true,"description":"Escalation context: situation description, affected scope (sprint/PI/project), quantified impact (delays, revenue, users), deadline or urgency, decision or action needed, options if applicable, supporting data (ticket counts, dates, metrics).\n"}] |
| outputs | [{"name":"memo","type":"text","description":"Complete escalation memo ready to send to leadership."}] |
| model_compatibility | ["claude","gpt-4","gemini","llama-3"] |
Craft Escalation Memo
Generate a structured escalation document for leadership. The memo must be evidence-based: every claim backed by data. Tone: professional, urgent without panic, factual.
When to Use
- A critical risk requires leadership attention or a decision beyond the team's scope
- A blocker is affecting delivery timeline and needs external intervention
- A scope, dependency, or capacity issue requires stakeholder alignment
- A risk assessment has flagged a Red or Critical Red item that needs escalation
Memo Structure
The memo MUST follow this structure. Each section has a specific purpose.
| Section | Purpose | Length |
|---|
| Situation | What is happening. Facts only, no speculation. | 2–4 sentences |
| Impact | Quantified consequences. Numbers, dates, scope. | 2–4 bullets |
| Urgency | Deadline or time sensitivity. Why act now. | 1–2 sentences |
| Ask | Specific decision or action needed from leadership. | 1–2 sentences |
| Options | If applicable, 2–3 alternatives with trade-offs. | Optional, 2–4 bullets |
Method
Step 1: Extract and validate evidence
From the provided escalation context, extract:
- Facts: What happened, when, who is affected
- Metrics: Ticket counts, days delayed, percentage impact, user/revenue numbers
- Dates: Deadlines, sprint/PI boundaries, when the issue was first observed
- Decisions needed: What exactly must be decided or approved
Rule: If a claim cannot be backed by data from the context, either omit it or mark it as "(unverified — needs confirmation)." Never fabricate numbers.
Step 2: Draft each section
Situation
- Lead with the core fact. No preamble.
- Use past tense for what has happened, present for current state.
- Avoid blame. Describe the situation, not who is at fault.
Impact
- Use bullets. Each bullet = one quantified consequence.
- Prefer numbers: "3 critical tickets blocked," "Release delayed by 5 days," "~2,000 users affected."
- If quantification is uncertain, use ranges: "Estimated 1–2 week delay."
Urgency
- State the deadline or trigger. "Sprint ends Friday." "PI planning in 5 days."
- Explain why waiting is costly. "Decision needed before we can commit to PI scope."
Ask
- One clear, actionable request. "Approve budget for contractor." "Confirm we can descope feature X." "Escalate to {team} for unblock."
- Avoid vague asks like "Please advise." Be specific.
Options
- Only include if there are genuine alternatives.
- For each option: brief description, trade-off, recommendation if clear.
- If there is only one path, omit this section.
Step 3: Apply tone rules
- Professional: No emotional language, no panic words ("disaster," "catastrophe")
- Urgent without panic: Convey seriousness through facts, not exclamation marks
- Factual: Every claim traces to the provided context
- Concise: Leadership has limited time. Every sentence earns its place.
Step 4: Validate against anti-patterns
See the anti-patterns table below. Rewrite any section that violates them.
Good vs Bad Escalation Framing
Good Situation:
- "The integration with {external system} has been blocked since {date}. The vendor has not responded to our last 3 escalation attempts. We have 2 critical tickets and 4 high-priority tickets dependent on this."
Bad Situation:
- "Things are really bad with the integration." (vague, emotional)
- "The vendor is incompetent and we're stuck." (blame, unprofessional)
Good Impact:
- "Release delayed by 5 days (from {date} to {date})"
- "3 critical tickets blocked, affecting ~40% of sprint scope"
- "Estimated revenue impact: {range} if not resolved by {date}"
Bad Impact:
- "This will cause major problems." (unquantified)
- "Everything is at risk." (vague)
Good Ask:
- "Request: Approve budget for a 2-week contractor to unblock the integration, or authorize descoping features X and Y from the release."
Bad Ask:
- "What should we do?" (vague, puts burden on reader)
- "We need help." (not actionable)
Anti-Patterns (NEVER Do These)
| Anti-Pattern | Why It's Bad | Do Instead |
|---|
| Vague impact ("major issues") | No basis for prioritization | Quantify: tickets, days, users, revenue |
| Blame or finger-pointing | Undermines collaboration | Describe the situation neutrally |
| Emotional language ("disaster") | Reduces credibility | Use factual, measured language |
| Unsupported claims | Leadership cannot act on speculation | Cite data or mark as unverified |
| Vague ask ("Please advise") | No clear next step | State specific decision or action needed |
| Burying the ask | Reader may miss it | Put the ask in a dedicated section |
| Excessive length | Leadership will skim or skip | Keep to 1 page; cut non-essential detail |
| Missing urgency | No reason to act now | State deadline and cost of delay |
Output Format
# Escalation Memo
**Subject**: {one-line summary}
**Date**: {date}
**From**: {role/team}
**To**: {audience}
---
## Situation
{Facts. What is happening. 2–4 sentences.}
## Impact
- {Quantified consequence 1}
- {Quantified consequence 2}
- {Quantified consequence 3}
## Urgency
{Deadline or trigger. Why act now. 1–2 sentences.}
## Ask
{Specific decision or action needed. 1–2 sentences.}
## Options (if applicable)
| Option | Description | Trade-off |
|--------|-------------|-----------|
| A | {brief} | {trade-off} |
| B | {brief} | {trade-off} |
**Recommendation**: {Option X} — {one-line reason}
---
## Supporting Data
{Optional: key metrics, ticket keys, dates for reference}
**Confidence**: {High | Medium | Low} — {one-line justification based on data completeness}
Self-Evaluation Checklist
Before delivering, verify:
If any check fails, revise before returning.
Error Handling
- Insufficient data: If the context lacks quantification, use "(estimated)" or "(unverified)" and recommend gathering more data before sending. Mark confidence as Low.
- No clear ask: If the context does not specify what decision is needed, add a placeholder: "[Ask to be defined: e.g., approve X, escalate to Y, descope Z]" and flag for human completion.
- Conflicting data: If the context contains contradictory information, note it: "Context contains conflicting data on {topic}. Recommend verification before sending."
Anti-Patterns
- NEVER write "This is urgent, fix now" without quantified impact. Every escalation must state what is at risk (SP, timeline, teams affected) and by when.
- NEVER assign blame to individuals. Escalation memos describe the situation and its impact, not who caused it.
- NEVER leave the Ask section vague. "Please help" is not an ask — "Approve option A by Wednesday" is.