postmortem

Extract numbered long-term memory candidates from the current task context.

1|Updated Apr 22, 2026

Search Tags:#context #evaluation #memory #postmortem #lessons #memory-management #recap

Ready to run

npx skills add https://github.com/terencefan/inority-workspace --skill postmortem-terencefan

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: postmortem
Source: https://github.com/terencefan/inority-workspace/tree/main/skills/postmortem
Command: npx skills add https://github.com/terencefan/inority-workspace --skill postmortem-terencefan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Extract reusable lessons from the current task context after debugging, execution, incident handling, runbook work, or visible rework. Use only when explicitly invoked as $postmortem and the goal is to synthesize numbered long-term-memory candidates from the current context, letting the user choose which numbers to keep and avoiding writing durable memory before selection.

Core Features & Use Cases

Promote the current context into a shortlist of reusable lessons that can be selectively memorized.
The output is a numbered list of durable memory candidates, chosen by the user.
Use cases include debugging wrap-ups, incident reviews, and post-runbook improvements.

Quick Start

Invoke the postmortem skill after completing a task to generate a numbered set of memory candidates and review them before deciding what to keep.

Frequently Asked Questions about postmortem

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract reusable lessons from debugging sessions or incident reviews?▼

Summarize your current context into a shortlist of reusable lessons by generating numbered long-term-memory candidates. You then review these postmortem candidates and select which specific items to keep before any durable write occurs.

What is the best way to save durable memory from task context without overwriting existing knowledge?▼

Generate a numbered set of memory candidates from your current context and manually select which numbers to keep. This prevents automatic durable writes, ensuring you only save verified lessons from debugging or runbook improvements without overwriting existing knowledge.

Does generating a postmortem recap automatically write lessons to long-term memory?▼

Generating a postmortem recap does not automatically write to long-term memory. It produces a numbered list of candidates from the current context, and the user must explicitly select which numbers to keep before any durable memory is written.

Can I use this approach to improve runbooks after execution or visible rework?▼

Yes, you can identify and extract actionable lessons from runbook execution or visible rework to form a reusable knowledge set. This synthesizes your current context into a shortlist of long-term-memory candidates for selective retention.

When should I avoid generating a postmortem recap for my current task?▼

Avoid generating a postmortem recap when explicit invocation is not intended or when the goal is not to synthesize numbered long-term-memory candidates. It requires explicit invocation and should not be used if immediate durable memory writing without selection is desired.