- Home
- Skills
- DevOps & Cloud
- incident postmortem writer skill
incident postmortem writer skill
Create evidence-backed, blameless incident postmortems from logs, alerts, and chat transcripts.
Free
Works with the AI Agents you already use
incident postmortem writer skill
Example session with this skill installed
I have an app.log and a chat transcript from yesterday's API outage between 14:00 and 15:30 UTC. Can you draft a blameless postmortem and build the timeline?
- Read your context and instructions
- Compiled the incident postmortem writer
I have processed the app.log and the chat transcript to build the incident report. I sanitized two internal IP addresses and three customer email addresses found in the logs.
Incident Postmortem: API Gateway Timeout (2026-10-04)
Date: 2026-10-04
Status: Draft
Severity: SEV-1
Summary
A misconfigured rate-limit policy triggered at 14:05 UTC, causing the API Gateway to drop 45% of inbound traffic. Service was restored at 15:12 UTC by reverting the gateway-config repository to the previous commit.
Impact
- Duration: 67 minutes.
- Affected Users: Approximately 12,000 requests failed with 429 or 503 status codes.
- Business Effect: Checkout flow was unavailable for users in the US-East region.
Timeline (UTC)
| Time | Source | Event | Status |
|---|---|---|---|
| 14:02:10 | Deploy | CI/CD merged PR #402 (Rate limit update) | Confirmed |
| 14:05:00 | app.log | First spikes of 429 Too Many Requests observed | Confirmed |
| 14:12:30 | Alert | PagerDuty: API Error Rate > 5% | Confirmed |
| 14:20:00 | Chat | On-call engineer (DevOps) joined bridge | Confirmed |
| 14:45:00 | Chat | Hypothesis: Policy v2_limit is overly restrictive | Inferred |
| 15:05:00 | Deploy | Rollback of PR #402 initiated | Confirmed |
| 15:12:00 | app.log | Error rates return to baseline | Confirmed |
Contributing Factors
- Trigger: Deployment of new rate-limiting logic with a syntax error in the threshold value.
- Vulnerability: The staging environment did not mirror production traffic volume, allowing the bug to pass CI.
- Factor: The alerting threshold was set high, delaying notification by 7 minutes after impact began.
Action Items
- Prevent: Add a dry-run mode to rate-limit deployments. (Owner: Platform Team, Due: 2026-10-20)
- Detect: Lower the API error rate alert threshold to 2%. (Owner: SRE, Due: 2026-10-10)
Next steps
- Review the "Inferred" timestamp at 14:45 with the responder to confirm the exact time the cause was identified.
- Assign a specific owner to the "dry-run mode" action item.
- Run the internal
lint_postmortem.pyscript once the final owners are added.
Connects securely to your tools. The creator never sees your data.
About this skill
When a system fails, the pressure to document what happened often leads to rushed reports, guesswork, and finger-pointing. This skill transforms raw incident data into objective, blameless postmortems that prioritize system learning over individual fault.
What it does
- Timeline construction merges logs, alerts, and chat transcripts into a single UTC-normalized sequence of events.
- Impact quantification calculates precise outage durations and identifies affected user segments based on evidence.
- Blameless analysis reframes human errors as systemic vulnerabilities, focusing on guardrails and process gaps.
- Automated linting scans drafts for blame-oriented language, missing sections, or weak action items.
- Data sanitization masks sensitive information like tokens, PII, and internal hostnames before final reporting.
How it works
- Intake where you provide logs, chat exports, and deploy history along with the incident window.
- Analysis using the included Python scripts to process timestamps and identify error patterns.
- Drafting the report using a structured template that separates triggers from contributing factors.
- Refinement via the linter to ensure the tone is professional and the action items have clear owners.
Frameworks & tools
This skill utilizes Python 3.9+ for log processing and timeline generation. It handles standard formats including ISO 8601, syslog, and JSON-line logs.
Why this beats prompting it yourself
Standard prompts often hallucinate missing details to "complete" a story, creating dangerous inaccuracies. This skill enforces strict evidence-based writing, marking unproven claims as Unknown and ensuring every action item includes a verification method.
Use cases
- Documenting production outages for internal engineering retrospectives.
- Generating sanitized incident summaries for customer-facing status pages.
- Performing root cause analysis (RCA) on near-miss events to improve system resilience.
- Auditing existing drafts to remove accusatory language and strengthen remediation plans.
Known limitations
The timeline builder only covers provided text sources and cannot process images or PDFs. Syslog-style logs without year headers default to the current year.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for DevOps & Cloud.
Timeline script only reads text sources. Log lines without years default to current year. Linter checks form and tone, not factual accuracy. Legal review required for regulated data.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account