retro

Synthesizes dated retros from the observability event log into committed archive files.

Updated Jun 28, 2026
One-click install
npx skills add https://github.com/sicambria/talkteach-asr --skill retro-sicambria
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: retro
Source: https://github.com/sicambria/talkteach-asr/tree/main/.claude/skills/retro
Command: npx skills add https://github.com/sicambria/talkteach-asr --skill retro-sicambria

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Decisions and events accumulate silently in the runtime event log (.harness/state/events.jsonl) and are never reviewed unless someone periodically turns them into a human-facing, committed artifact. This Skill closes that loop by reading the log and writing a dated retro summarizing what happened, notable decisions, and follow-ups. ## Core Features & Use Cases - Event Log Summarization: Runs the events CLI with --summarize to pull counts, event types, and timestamps without hand-parsing JSONL. - Structured Retro Generation: Writes a dated Markdown retro under .harness/archive/retros/ covering what happened, decisions logged, notable patterns, and concrete follow-ups. - Loop-Closing Logging: Logs the retro itself as an event so the dashboard's retro links stay current. - Use Case: At the end of a work session or sprint, ask for a retro to capture decisions and gate overrides while they are still cheap to reconstruct, producing a committed artifact for the team. ## Quick Start Run a retro summarizing everything logged since the last one and write it to the retros archive.

Frequently Asked Questions about retro

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

FAQPage Schema
How do I generate a retro from an event log?▼

Run the events CLI with --summarize --events to get the full event list and counts, then synthesize a narrative covering what happened, decisions logged, notable patterns, and follow-ups. Write the result as a dated Markdown file under .harness/archive/retros/.

What is the difference between a retro and a postmortem?▼

A retro summarizes routine session activity from the event log, while a postmortem documents incidents. Both close the loop with a committed artifact, but retros are periodic and event-grounded rather than incident-driven.

When should I not run a retro?▼

Do not run a retro when the event log has no new events since the last retro. The Skill instructs you to say so plainly rather than manufacturing content from an empty window.

Does the retro skill require parsing JSONL manually?▼

No. The events.mjs script provides a --summarize flag that returns total counts, events grouped by type, first and last timestamps, and the full event list, so no hand-parsing of the JSONL file is needed.

How does the retro stay linked to the dashboard?▼

After writing the retro file, the Skill logs it as a retro-type event via events.mjs. The materialize script's dashboard reads these events to populate its retro links section.