skill-improver

Analyze run artifacts to propose concrete skill and workflow edits.

339|58|Updated Jan 10, 2026

Search Tags:#logs #openspec #workflows #analyze #improve #retrospect #run-artifacts

Ready to run

npx skills add https://github.com/Heyvhuang/ship-faster --skill skill-improver-heyvhuang

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-improver
Source: https://github.com/Heyvhuang/ship-faster/tree/main/skills/skill-improver
Command: npx skills add https://github.com/Heyvhuang/ship-faster --skill skill-improver-heyvhuang

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Retrospect and improve Skills/Workflows based on real execution artifacts, making the skill chain more stable with use.

Core Features & Use Cases

Analyze run artifacts (events.jsonl and state.json) from runs/ to identify failures, bottlenecks, unclear prompts, and missing I/O contracts.
Propose concrete edits to skills and workflows, producing an improvements.md and optional patch to SKILL.md for deterministic updates.
Use case: When a skill underperforms across multiple runs, generate a prioritized list of changes to improve reliability and traceability.

Quick Start

Run the improver against a runs/ directory (and/or openspec/changes/) to produce improvements.md and an optional SKILL.md patch; review the suggested changes and apply as tasks.
If you want to see concrete edits, inspect the generated improvements.md.

Frequently Asked Questions about skill-improver

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

FAQPage Schema

How do I analyze run artifacts to improve skill reliability?▼

To analyze run artifacts, point the skill at a runs/ directory to process state.json and events.jsonl. It identifies failures, bottlenecks, and unclear prompts, then generates an improvements.md file with concrete edits to increase skill reliability.

What is the best way to retrospect on OpenSpec changes and workflow failures?▼

The best way to retrospect on OpenSpec changes is to analyze logs and run artifacts across multiple runs. This identifies failures and missing I/O contracts, proposing concrete edits to workflows for deterministic updates and better traceability.

Does this skill automatically patch SKILL.md files based on logs?▼

Yes, the skill can optionally patch SKILL.md files based on log analysis. It reads events.jsonl and state.json to propose concrete edits, outputting an optional patch alongside an improvements.md file for deterministic updates.

How do I identify unclear prompts and I/O gaps from execution logs?▼

You can identify unclear prompts and I/O gaps by running the skill against your execution logs. It reads events.jsonl and state.json from your runs/ directory to pinpoint these issues and propose prioritized changes in improvements.md.

Can I use this to fix workflows that fail across multiple runs?▼

Yes, you can use this to fix workflows that fail across multiple runs. It analyzes the run artifacts to find recurring failures and bottlenecks, generating a prioritized list of changes in improvements.md to make the skill chain more stable.