Knowledge capture skill

Project retrospective

Generate LESSONS.md files that capture institutional knowledge, especially failures. Think like a journalist writing about your own project.

What this skill does

Creates structured retrospective documents that capture what worked, what didn't, and—most importantly—what you thought you were building versus what was actually needed. The goal is honest, specific documentation that saves others from the same mistakes.

📋

Investigations

Research projects

📅

Events

Conferences, workshops

📰

Publications

Newsletters, podcasts

🛠️

Tools

Newsroom software

The most important section: "The real problem"

This is the most valuable part of any retrospective. It answers:

"What did we THINK we were building vs. what was ACTUALLY needed?"

Good example

"We built a comprehensive tagging system when users just needed full-text search. Three weeks on features no one used."

Bad example (too generic)

"We learned the importance of user research."

LESSONS.md template structure

1

Project metadata

Name, dates, status, author

2

Summary

What it did, impact, why it matters

3

What worked

Technical wins, process wins

4

What didn't work

Critical failures, technical debt, external factors

5

The real problem

What you thought vs. what was needed

6

Recommendations

If continuing, if starting fresh, tech stack verdict

7

Reusable artifacts

Components others can use

8

Questions for next time

Unanswered questions worth investigating

The specificity test

For each item in "What didn't work," ask: Can I name the specific decision? Can I quantify the impact? Would this help someone avoid the same mistake?

Too vague

  • Communication could have been better
  • We underestimated the complexity
  • Testing was insufficient

Specific and actionable

  • Skipped schema validation on data files. Cost: 3 hours debugging a typo that caused silent failures.
  • Built custom date picker when browser native input would have worked. 2 days wasted.
  • No error messages when data fails to load—users just see blank screen.

Red flags in your writing

If you find yourself writing these, stop and be more specific. These are placeholders for real insights:

"Communication is key"
"We learned the importance of..."
"Going forward, we should..."
"Challenges included..."
"There were some issues with..."

A strong "real problem" example

The gap between assumption and reality

"We built an admin dashboard for editors when they actually needed a Slack bot. They live in Slack—forcing them to open a web app was friction they'd never accept. The dashboard has 2 monthly active users; the Slack bot prototype we built in a day has 47."

Journalism-specific templates

research-project.md

Investigations, data journalism

For deep-dive investigative projects with data analysis, source management, and publication timelines.

event.md

Conferences, workshops

For in-person and virtual events with attendance metrics, speaker feedback, and logistics lessons.

publication.md

Newsletters, podcasts

For ongoing content series with audience growth, production workflows, and sustainability considerations.

editorial-tool.md

Newsroom software, AI tools

For internal tools with adoption metrics, integration challenges, and user feedback tracking.

Voice guidelines

  • Honest, specific, slightly self-deprecating — Like explaining to a friend why the project took twice as long
  • No corporate speak or blame-shifting — Name specific mistakes, not vague "challenges"
  • Quantify when possible — "2 days wasted" is better than "some time lost"
  • Focus on decisions, not people — Include failures with context, exclude blame for individuals

Installation

# Clone the repository

git clone https://github.com/jamditis/claude-skills-journalism.git

# Copy the skill to your Claude config

cp -r claude-skills-journalism/project-retrospective ~/.claude/skills/

Or download just this skill from the GitHub repository.

Start capturing lessons

The best retrospectives are written by people who got burned and want to save others from the same fate.

View on GitHub