- Home
- Skills
- DevOps & Cloud
- Postmortem Writer
Postmortem Writer
Transforms incident logs and alert timelines into blameless, evidence-backed postmortem reports.
$39
Works with the AI tools you already use
Postmortem Writer
Example session with this skill installed
Draft a postmortem for the API outage yesterday. Here is the Slack export from #incident-api-104 and the PagerDuty alert timeline. Focus on why the auto-scaler failed to trigger.
- Read your context and instructions
- Compiled the postmortem writer
- Generated the document
I have analyzed the Slack transcript and PagerDuty events. I normalized 42 events into a UTC timeline and identified a 12-minute gap between the first 5xx spikes and the initial page.
Incident Report: API Scaling Failure (SEV-2)
Summary: The production API experienced 100% error rates for 18 minutes due to an auto-scaler configuration lockout.
| Metric | Value |
|---|---|
| MTTD | 4m 12s |
| MTTA | 2m 45s |
| MTTR | 18m 22s |
| Severity | SEV-2 |
Timeline Summary
- 2023-10-12 14:02:10 UTC: First 503 errors detected in
ap-southeast-2. - 2023-10-12 14:06:22 UTC: PagerDuty alert #882 fired (High Urgency).
- 2023-10-12 14:09:07 UTC: On-call engineer acknowledged alert.
- 2023-10-12 14:15:30 UTC: Root cause identified: Terraform lock held by stale CI job prevented scaling.
- 2023-10-12 14:20:32 UTC: Manual lock release; scaling resumed.
Root Cause Analysis
- Trigger: Sudden traffic spike exceeded current node capacity.
Root Cause: A failed CI pipeline left a global state lock on the production workspace, preventing the auto-scaler from updating the desired instance count.
Contributing Factor: The auto-scaler's health check did not alert when Terraform operations failed due to locking.
Action Items
- Prevents: Implement 10-minute TTL on Terraform state locks. (Owner: Platform Team)
- Detects: Add CloudWatch alarm for
ScalingFailedevents in the provider log. (Owner: SRE) - Mitigate: Document manual state unlock procedure in the emergency runbook. (Owner: On-call)
Next steps
- Review the 4-minute gap between error spikes and the PagerDuty alert; the threshold may be too high.
- Assign owners and due dates to the action items above.
- Verify if the stale CI job (13:45 UTC) is a recurring failure.
postmortem-writerwkr.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
What you get
About this skill
The problem
Engineering teams lose hours manually reconstructing incident timelines from Slack logs and PagerDuty alerts. These documents often suffer from subjective bias, blame-coded language, and metrics that don't actually match the evidence.
What it does
- Normalizes and merges timestamps from Slack, PagerDuty, and log files into a single UTC timeline.
- Calculates precise MTTD, MTTA, and MTTR metrics directly from tagged incident events.
- Identifies the technical trigger, root cause, and contributing factors using blameless 5-Whys analysis.
- Generates a structured Markdown report with categorized action items (prevents, detects, mitigates).
- Flags timeline gaps and unsupported causal claims for human verification.
Frameworks & tools
Python-based scripts for timeline parsing and report rendering. Supports JSON exports from Slack, PagerDuty, and Opsgenie, plus CSV and plain text logs.
Why this beats prompting it yourself
Generic LLMs often hallucinate timestamps or "guess" root causes when data is missing. This skill enforces strict evidence-based drafting, ensuring every claim in the report is backed by a specific entry in the normalized timeline while strictly maintaining a blameless engineering culture.
Use cases
- Drafting SEV1/SEV2 root cause analysis documents after a system outage.
- Auditing human-written postmortems for timeline gaps or blame-heavy language.
- Converting messy incident Slack channels into readable executive summaries.
- Generating standardized incident metrics for quarterly reliability reviews.
Known limitations
Does not connect directly to external APIs. Users must provide text or JSON exports of incident data. Does not perform automated code-level debugging.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean 5 days ago
- Passed all security checks, Safe to install