retro

Conducts a retrospective on a coding session to suggest agent environment improvements.

Updated Aug 22, 2026
One-click install
npx skills add https://github.com/MSC72m/DevForge --skill retro-msc72m
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: retro
Source: https://github.com/MSC72m/DevForge/tree/main/skills/retro
Command: npx skills add https://github.com/MSC72m/DevForge --skill retro-msc72m

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After a coding session, it is hard to know why an AI agent struggled, made mistakes, or wasted tokens. This Skill analyzes session logs and turns those friction points into concrete improvements for the agent's environment, such as better navigation pointers, automated checks, and coding standards. ## Core Features & Use Cases - Session Log Analysis: Reads primary sources from a specified or current coding session to find improvement candidates. - Categorized Findings: Evaluates navigation, automated checks, coding standards, AGENTS.md bloat, tool economy, no-op instructions, and information access, then ranks findings by severity. - Guardrail-First Recommendations: Classifies violations as mechanical (fixed with deterministic checks like linters, pre-commit hooks, or CI jobs) versus judgement calls (documented in CODING_STANDARDS.md). - Use Case: After an agent repeatedly failed to find a config file and shipped un-linted code, run a retrospective to add a navigation pointer to AGENTS.md and wire the repo's existing lint script into CI. ## Quick Start Ask the agent to run a retrospective on the current coding session and suggest environment improvements.

Frequently Asked Questions about retro

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

FAQPage Schema
How do I run a retrospective on an AI coding session?▼

Invoke the retro skill and specify the session to analyze, or let it default to the current session. It reads the session logs, identifies improvement candidates across seven categories, and presents them ranked by severity.

What kinds of improvements does a coding session retrospective suggest?▼

It suggests navigation pointers for hard-to-find files, automated checks like linting and tests, coding standards rules for the reviewer agent, AGENTS.md cleanup, tool economy fixes, removal of no-op instructions, and better information access such as teed dev server logs.

When should a coding standard become an automated check instead of a written rule?▼

Mechanical violations such as banned APIs, import shapes, or file-location rules should become deterministic checks in a linter, pre-commit hook, or CI job. Written standards in CODING_STANDARDS.md are reserved for genuine judgement calls like cross-file style consistency.

Does the retro skill modify code or just make recommendations?▼

It only analyzes and presents improvement candidates to the user in order of severity. The user decides which suggestions to implement, and the skill is configured to disallow implicit model invocation.

Why should the reviewer agent enforce coding standards instead of the implementation agent?▼

The implementation agent has the most context pressure from exploration, coding, and debugging, while the reviewer receives only a diff. Assigning standards enforcement to review keeps the implementation context lean and focused.