Skill Optimizer

Trust Score 88/100

Diagnose and optimize Agent Skills (SKILL.md) using session transcripts and static analysis to improve triggers, workflows, and token efficiency.

triggers:optimize skillaudit skillsskill optimizerdiagnose skillskill audit

GitHub SKILL.md

What it does

Runs an evidence-driven audit of Agent Skills using historical session data and static checks. Measures trigger rates, post-invocation user reactions, workflow completion, static quality rules, undertrigger/overtrigger rates, environment consistency, and token economics to produce prioritized fixes.

When to use it

Use when skills are misfiring, underused, or you need a data-driven plan to improve skill routing and descriptions. Good for library maintenance, onboarding new skills, or periodic quality audits.

What's included

Scripts: the repository contains analysis scripts to scan Claude/Codex session JSONL files and compute the eight diagnostic dimensions (fetch output indicates has_scripts=false for this bundle, but the skill describes scanning scripts).
References: guidance and research-backed rationale for suggested description rewrites and prioritization.
Instructions: step-by-step workflow for collecting session data, running all eight analysis dimensions, and producing a scored report with P0/P1/P2 fixes.

Compatible agents

Targets agents and environments where session transcripts are available (Claude Code, Codex) and where read-only auditing of skill files is permitted.

Audit Summary

A comprehensive skill optimization/audit tool that analyzes SKILL.md files and session transcripts across 8 dimensions including trigger rate, user reaction, workflow completion, static quality, undertrigger detection, cross-skill conflicts, environment consistency, and token economics. Well-structured with clear workflow steps, research-backed rationale, and detailed scoring rubrics. No bundled scripts — relies on agent executing bash/python commands directly, which limits reproducibility.

Watch Out

No bundled scripts — relies on agent ad-hoc execution which may vary across runs
Requires access to session JSONL files (~/.claude/projects or ~/.codex/sessions) which may not exist in all environments
Very long SKILL.md (~900+ words) — ironic for a skill that advocates <500 words and progressive disclosure

Notes

Read-only approach (never modifies skill files) is a strong security positive. Well-researched with academic citations. The 8-dimension analysis framework is thorough and well-structured. Main weaknesses: no scripts for reproducible analysis, SKILL.md itself violates its own word-count/progressive-disclosure recommendations, and without session data available the skill degrades to static-only analysis which is less valuable.

Information

Repository: opencode-skills-collection
Stars: 23
Installs: 0

Trust Score

Overall88

Security95

Code Quality82

Architecture78

Usefulness72

More from opencode-skills-collection

Python Performance Optimization

Profile, analyze, and optimize Python applications for CPU and memory efficiency using profiling tools and performance best practices.

API Security Testing Workflow

Structured workflow for testing REST and GraphQL API security: auth, authorization, rate limiting, input validation and error handling.

Azure AI Projects SDK (TypeScript)

TypeScript SDK and examples for managing Azure AI Projects: agents, connections, deployments, datasets, indexes, and evaluations.

Memory Forensics Playbook

Practical guidance and commands for acquiring, analyzing, and extracting artifacts from memory dumps using tools like Volatility3 and common acquisition methods

Testing Patterns & Utilities

Guidelines and utilities for TDD, factory-based test data, mocking strategies, and testing patterns for React/TypeScript projects.

Reverse Engineer

Guided methodology and best practices for binary reverse engineering, covering static and dynamic analysis workflows and common tooling.

Related Skills

Readwise Reader Document Management

Manage Readwise Reader documents: list, save, search, move, tag, highlight, export and bulk-edit via official and custom CLIs.

ezBookkeeping API Tools

Command-line API tools for ezBookkeeping: record and query transactions, retrieve accounts/categories/tags, and fetch exchange rates for self-hosted personal fi

Kickbox Automation (Composio / Rube MCP)

Automate Kickbox workflows via the Composio Kickbox toolkit on a Rube MCP server: discover tools, manage connections, and execute schema-compliant tool calls.

Tmux Bridge

Give an agent controlled access to a local tmux terminal bridge so it can run shell commands, capture output, and manage sessions on the user's machine.

ArcKit Architecture Conformance

Run a systematic architecture conformance assessment that compares ADRs and principles against HLD/DLD artifacts to surface drift, violations, and technical deb

SourceSage CLI

Generate concise, LLM-friendly repository summaries (Markdown) using the SourceSage CLI — supports lite mode, language switching, and targeted repo analysis.

Nit (Nitter CLI)

Terminal client to browse tweets, view profiles, and search posts via Nitter instances without opening a browser.

hn — Hacker News CLI

Terminal-based Hacker News client with a CLI and optional interactive TUI for browsing top/new/best/ask/show/jobs stories, viewing threads with nested comments,

Back to Skills