SKILL.md
PulseAugur coverage of SKILL.md — every cluster mentioning SKILL.md across labs, papers, and developer communities, ranked by signal.
8 day(s) with sentiment data
-
SimpleEnglish agent reduces LLM reply defects by 86%
The SimpleEnglish agent skill, designed to enforce Simplified Technical English (ASD-STE100) standards in LLM outputs, has reported a significant reduction in visible reply defects. An audit revealed that while previous…
-
AI agent skill files offer structured workflows for repeatable tasks
This guide explains how to create AI agent skill files, which provide structured workflows for handling recurring tasks, differentiating them from single-use prompts. A skill file defines the trigger, goal, scope, workf…
-
Claude Code's /skill-doctor tool identifies unused skills and token waste
Claude Code has introduced a new diagnostic tool called /skill-doctor, which reports on the token cost and usage frequency of individual skills loaded into the system. The tool revealed that a significant portion of the…
-
AI Agent Skills: Anatomy of a Debugging Method for Longevity
Ramdai Bista has detailed a method for creating "SKILL.md" files that remain effective for extended periods, particularly in AI agent applications. The approach focuses on systematic debugging techniques to address bugs…
-
Claude Code plugin streamlines Notion session summaries with new features
A developer has released a new plugin for Claude Code called "cseeman/claude-journal" that enhances session summarization within Notion. The plugin, an update to a previous script, now creates structured entries with he…
-
AI Agents: Distinguishing Skills, MCP, Function Calling, and A2A
The article clarifies the distinctions between Agent Skills, Model Context Protocol (MCP), Function Calling, and Agent-to-Agent (A2A) communication in AI development. Agent Skills are described as self-contained folders…
-
Claude AI assistant compromised via malicious link within chatbot
A user was compromised through the Claude AI assistant when a malicious link was delivered directly within the chatbot interface. The user intended to install an audio transcription application and received a link from …
-
Claude skills corpus reveals widespread file duplication
The author discovered that a significant portion of files labeled as "skills" within a Claude corpus were actually backups, not functional skills. This finding emerged during an attempt to measure duplication across the…
-
Malware hijacks Claude sessions, draining accounts; new AI agent supply-chain risks emerge
Malware targeting user machines has led to the hijacking of active Claude login sessions, allowing attackers to drain paid usage quotas. Anthropic responded by forcing users to log out, removing saved payment methods, a…
-
Anthropic's Agent Skills offer reusable instructions for AI models
Anthropic has introduced Agent Skills, a feature designed to provide AI models with structured, reusable instructions and resources for specific tasks. These skills are essentially folders containing a SKILL.md file, wh…
-
Developer fixes 'skill' name collision bug in Vodou system
A developer encountered a bug in the Vodou system where two distinct entities, a file-based SKILL.md folder and a prompt template, shared the same name 'skill'. This ambiguity led to incorrect data retrieval, as queries…
-
Anthropic's Claude Cowork adds 'Record a Skill' for AI workflow automation
Anthropic has introduced a new feature called "Record a Skill" within its Claude Cowork application, enabling users to demonstrate tasks through screen recording, voice narration, and interaction capture. This feature d…
-
Anthropic launches Claude Skills API for custom skill management
Anthropic has officially launched the Claude Skills API, allowing developers to create, version, and manage custom skills directly through the API without needing to host their own infrastructure. This feature, now gene…
-
Developer creates SkillEval to unit test LLM agent prompts
A developer has created a testing framework called SkillEval to address the lack of rigorous testing for agent skills, which are essentially prompts rather than traditional code. This tool allows developers to run agent…
-
AI agent skills need rigorous testing beyond single successful runs
Testing AI agent skills requires more than a single successful run, as a skill that works once but fails repeatedly is worse than no skill at all. Effective testing involves evaluating the skill's performance across var…
-
Qwen Code 0.21.11 adds Agent Plugins v1 support, with caveats
Qwen Code version 0.21.11 introduces native support for Agent Plugins v1, allowing users to install and link these packages without needing to convert their manifest files. However, not all components of an Agent Plugin…
-
AI Agent Skill Loading Debugging Checklist Revealed
This article discusses a specific debugging issue encountered when developing AI agents, particularly concerning the loading of agent skills. The author, Ramdai Bista, highlights a silent failure mode where a skill migh…
-
AI Agent Skills Pose Security Risks, Audit Recommended
The article warns that agent skills, which are code snippets used by AI agents, should be treated as untrusted code. It suggests a three-step audit process before installing any such skills to mitigate security risks. T…
-
Claude Code: Top 6 Skills for Developers Recommended
A guide recommends six specific Claude Code skills for developers, emphasizing the importance of installing only one skill per development phase to avoid unpredictable behavior. The selection criteria prioritize officia…
-
Anthropic's Claude limits skill context to 500 lines
Anthropic's Claude AI model has a context window limitation that requires users to keep their "Claude Code SKILL.md" files under 500 lines. Bloated skills consume context for the entire session, impacting performance. D…