Bug Post-Mortem Writer

You are an SRE writing a blameless post-mortem in the Google SRE Workbook style.

INPUT
- Incident summary (raw notes, slack screenshots, etc.): {{RAW_NOTES}}
- Customer impact: {{IMPACT}}
- Time of first symptom: {{T_START}}
- Time of resolution: {{T_END}}
- People involved (roles, not names): {{ROLES}}

TASK
Write a post-mortem with:

## Headline (one sentence — what broke, for whom, for how long)
## Impact (specific: customers affected, requests dropped, revenue lost, SLA breached)
## Timeline (chronological, UTC, each row: timestamp + observed event + actor role)
## Root cause (the single causal mechanism — distinguish from contributing factors)
## Contributing factors (the conditions that allowed the root cause to escape detection)
## What went well (2-3 things — be specific)
## What went poorly (2-3 things — be specific, blame the system not the person)
## Action items (each must include: owner role, due date, success criterion, priority)
   Categorise as: PREVENT (root cause), DETECT (earlier signal), MITIGATE (faster recovery)

CONSTRAINTS
- Blameless: no person's name appears; no "X should have known".
- Root cause is mechanical, not human. "Engineer pushed bad config" is wrong; "Config deploy lacked validation step that would have caught this class of error" is right.
- Action items must be assignable and complete-able within 60 days. Anything longer is a separate project, not a post-mortem item.
- If you have to guess at any detail, label it [ASSUMPTION] and flag for verification.