Agent Skills look simple, but they are one of the most powerful building blocks in modern AI coding workflows. In this episode, Kaushik and Iury break down when to use skills, how progressive disclosure works, and how skills compare with commands, instructions, and MCPs.
Full shownotes at fragmentedpodcast.com.
Show Notes #
Main References #
- Progressive Disclosure - how skills are loaded into context
- Agent Skills Open Specification
- AAIF (Agentic AI Foundation) - Linux Foundation initiative for AI interoperability
- Needle in a Haystack Problem - original “Lost in the Middle” paper
- Agent-Invokable vs User-Invokable - merging skills and commands in Claude Code
Creating Skills #
- Skill Creator - Anthropic’s skill for creating new agent skills
- Claude Code frontmatter reference
- see
model: *&context: fork
- see
Using other Skills #
- Anthropic Skills GitHub Repository - official collection of Claude skills and examples
- Clawdhub - Clawdbot’s skill hub. All versions are archived here
- SKILLS.sh - Vercel’s skills hub
Warnings before installing random skills #
Don’t install from hubs blindly.
Inspect the repo code before adding anything to your agent.
- Prompt Injection Attacks - OWASP guide to LLM prompt injection vulnerabilities
- OpenClaw <- MoltBot <- Clawdbot
- OpenClaw Security Analysis - analysis of prompt injection risks in open agent frameworks
- Malware found in a top-downloaded Clawhub skill - incident report thread
Additional resources #
- Few-Shot Prompting - improving outputs with examples
- .agents/skills - proposal to standardize the skills folder path
- Vercel: AGENTS.md vs Skills - comparison of agent instruction methods
Get in touch #
We’d love to hear from you. Email is the best way to reach us or you can check our contact page for other ways.
We want to hear all the feedback: what’s working, what’s not, topics you’d like to hear more on.
Co-hosts: #
We transitioned from Android development to AI starting with
Ep. #300. Listen to that episode for the full story behind our new direction.