agent-skills
agent-skills copied to clipboard
addyosmani
Production-grade engineering skills for AI coding agents.
## Summary fix: resolve #122 — /build不能让agent触发git ## Problem **Severity**: `High` | **File**: `.claude/commands/build.md` The build command should automatically trigger the git worktree flow instead of requiring manual activation. This...
## Summary fix: resolve #124 — Question: CI validation for skills? ## Problem **Severity**: `Medium` | **File**: `skills/VALIDATION.md` Create documentation for a CI validation pattern that allows skills to define...
## Summary fix: resolve #146 — Скилы ## Problem **Severity**: `Medium` | **File**: `skills/README.md` The repository lacks a central README file explaining how to use, create, and manage skills. This...
## Summary fix: resolve #150 — Recommended way to handle V2 features without polluting V1 spec? ## Problem **Severity**: `Medium` | **File**: `.claude/commands/spec.md` The spec.md file should include guidance on...
`.claude-plugin/plugin.json` declares `"version": "1.0.0"` (what the Claude marketplace shows), but the latest GitHub release and git tag is `0.6.0`. With multiple variants in the repo (Claude, Gemini, OpenCode), I wasn't...
Several SKILL.md files contain script references in their "Usage" or "How It Works" sections that do not exist in the repository. ## Example \`spec-driven-development/SKILL.md\` references: \`\`\`bash /mnt/skills/user/spec-driven-development/scripts/fetch-spec.sh [args] \`\`\` But...
Hermes Agent (https://hermes-agent.nousresearch.com/) is a capable Python-based AI agent platform that supports SKILL.md-based skills. It would benefit from integration documentation in this repository. ## Proposal 1. Add Hermes Agent to...
During /spec I have lots of ideas, many of them I need to leave behind for the V1. 1. Should I mention in the spec that this is a V1,...
## 🎯 Objective Expands the core agent skills library from 20 to 25 production-grade workflows. These additions are directly ported from the advanced **AgentOps Cockpit** governance engine, bringing state-of-the-art autonomous...
## Summary Adds a new top-level skill, `prompt-injection-defense`, covering OWASP LLM01 across the inputs an agent reads rather than the code it writes. The skill treats prompt injection as a...
Metadata
Owner
Metadata
Production-grade engineering skills for AI coding agents.