retro

Conduct retrospectives on product or experiment loops and update state files.

1|Updated Feb 20, 2026
One-click install
npx skills add https://github.com/october-academy/agnt --skill retro-october-academy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: retro
Source: https://github.com/october-academy/agnt/tree/main/skills/retro
Command: npx skills add https://github.com/october-academy/agnt --skill retro-october-academy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Helps teams and solo builders perform structured retrospectives on an experiment or product loop by collecting context, prompting targeted reflection, and producing an actionable next-actions summary without manual state bookkeeping.

Core Features & Use Cases

  • Automated context collection: Reads AGNT_DIR state.json and journey-brief.md to extract verdicts, offers, and launch plans.
  • Guided reflection and branching: Prompts the user with targeted questions, handles verdict branches (CONTINUE/PIVOT/KILL/TOO EARLY), and records chosen next actions.
  • State management and integrations: Updates state.json and journey-brief.md, applies defaulting rules, performs KILL→A artifact resets when requested, and attempts MCP submit_practice or queues a pending sync event if the tool is unavailable.

Quick Start

Start a retrospective loop by invoking the retro workflow, answer the guided questions, and confirm the Next Actions to update state and journey-brief.

Frequently Asked Questions about retro

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

FAQPage Schema
How do I conduct a structured retrospective for an experiment loop?

A structured retrospective for an experiment loop is conducted by reading state.json and journey-brief.md, prompting targeted reflection questions, and summarizing verdicts into actionable next steps. The process handles CONTINUE, PIVOT, and KILL branching while automatically updating your state files.

How do I update state.json and journey-brief.md after a product loop ends?

You can update state.json and journey-brief.md after a product loop by running an automated retrospective workflow that applies default schema rules, increments loops_completed, and records chosen next actions. If a KILL verdict is selected, it performs an artifact reset to clean the directory.

Can I submit retrospective learnings to MCP if the integration tool is unavailable?

If the MCP submit_practice tool is unavailable, the retrospective workflow queues a pending sync event locally instead of failing. This ensures your learnings are captured and can be submitted to MCP automatically once the tool becomes accessible again.

What is the best way to handle CONTINUE, PIVOT, or KILL verdicts in agile state management?

Handling CONTINUE, PIVOT, or KILL verdicts in agile state management is best done through guided reflection that branches execution logic based on the chosen verdict. The workflow records the decision, updates Next Actions, and triggers a KILL-to-A artifact reset when a KILL scenario is selected.

When do I need to reset artifacts in a product retrospective workflow?

Artifacts need to be reset in a product retrospective workflow when a KILL verdict is chosen during the guided reflection. This KILL-to-A artifact reset cleans the state configuration, allowing you to start a fresh experiment loop without leftover configuration conflicts.

Does the retro workflow require manual state bookkeeping for journey-brief updates?

No, the retro workflow does not require manual state bookkeeping for journey-brief updates. It automatically reads existing context, enforces default schema rules, increments loops_completed, and writes the updated Next Actions directly to state.json and journey-brief.md.