Topic hub
Agent skills
Portable skill files, behavioral contracts, tool instructions, and evals that make agent capability reusable.
Retrieval answer
Portable skill files, behavioral contracts, tool instructions, and evals that make agent capability reusable. Agent skills package repeatable behavior outside the current chat. The important boundary is whether the skill changes task outcomes, not whether the instruction prose looks good. Skills become infrastructure when they are versioned, evaluated, and shared across agents.
Pattern memory
What patterns are emerging?
- medium
Skills become a portable capability layer
Agent skills are emerging as a portable capability layer, but their value depends on progressive disclosure, provenance, security review, and behavioral evaluation.
Field notes
What should readers understand next?
Dr. Skill Audits What An Agent Loads Before It Works
Dr. Skill scans skills and MCP servers for collisions, duplication, secrets, drift, missing metadata, and unused loadout, with local and CI-friendly commands.
Genkit Adds Progressive Disclosure For Agent Skills
Genkit now loads Agent Skills through middleware that discovers SKILL.md metadata first and activates full instructions, references, and scripts only when needed.
Mistral Treats Prompts And Skills As Production Records
Mistral Studio adds immutable versions, ownership, promotion labels, lineage, rollback, and audit logs for prompts and skills used in production AI systems.
Stripe Builds A Shared Agent Platform For Knowledge Work
Stripe's Kai combines surface-agnostic APIs, domain-owned AgentStudio assets, per-session sandboxes, long-horizon state, and more than 1,000 internal tools and skills.
Claude Cowork Turns Screen Recordings Into Reusable Skills
Claude Cowork's recorded-skill flow makes workflow capture a first-party path from human demonstration to reusable agent capability.
Hermes Agent Makes the Learning Loop Part of the Runtime
Nous Research's Hermes Agent packages memory, skills, messaging, scheduling, tool use, and sandboxing as a runtime rather than a single chat surface.
Raw signals
What changed recently?
Agent skills need behavioral evals, not prose review
Benchmarks show that expert-authored skills can help while self-generated skills can underperform a no-skill baseline.
Improve routes architecture and execution to different models
The Improve skill audits a repository with a stronger planning model and delegates bounded fixes to cheaper executors.
Large-task planning moves from chat into a ticket graph
Matt Pocock's workflow turns discovery, specification, tickets, implementation, and review into durable linked artifacts.
Animation Vocabulary Turns Motion References into an Agent Skill
A compact design skill maps visual intent to named motion patterns, giving interface agents a reusable language for implementing animation decisions.
A GitHub Skill Can Resolve Review Comments as a Batch
A reusable Codex workflow gathers pull-request feedback, plans changes, applies fixes, and reports resolution state as one auditable review task.
Pydantic AI loads capabilities only when needed
A capability can bundle instructions, tools, model settings, and hooks while exposing only a compact description until activation.
Creative Ideation Becomes a Routable Agent Skill
A packaged workflow decomposes creative exploration into reference gathering, divergence, critique, and selection instead of asking a model to be creative in one step.
Anthropic: Goal-Scoped Agent Loops
The archive captures X source as a dated public record from X source. It documents long-running work gaining explicit goals, state, stopping rules, and recovery and is retained as pressure-testing evidence for the goal-scoped agent loops trend.
Agentic Resource Discovery Indexes Skills, MCP, and Agents
Hugging Face treats agent capabilities as searchable resources, creating a discovery layer for tools that can be loaded at runtime.
Hermes Turns Solved Tasks into New Skills
An agent runtime can preserve successful procedures as reusable capability modules, converting execution traces into a compounding operational library.
CLAUDE.md Shrinks as Skills, Hooks, and Subagents Mature
Agent behavior moves out of one giant context file into scoped mechanisms that load only for the relevant task and lifecycle stage.
Codex Turns a Recorded Mac Workflow into a Reusable Skill
Record and Replay captures a human procedure, converts it into agent instructions, and preserves the successful interaction as reusable behavior.
Taste Lab Uses a Design System as a Reasoning Brief
Taste Lab packages aesthetic decisions and implementation rules as agent-readable context rather than relying on vague style prompts.
Last30Days Packages Social Research as an Agent Skill
A reusable research workflow gathers recent public discussion, filters repetition, and produces a time-bounded evidence set for another agent or analyst.
Claude Code: Verification Bandwidth
The archive captures Claude Code as a dated public record from Claude Code. It documents evaluation, review, and observability becoming the bottleneck after generation accelerates and is retained as pressure-testing evidence for the verification bandwidth trend.
GitHub / cursor/plugins: Verification Bandwidth
The archive captures GitHub / cursor/plugins as a dated public record from GitHub / cursor/plugins. It documents evaluation, review, and observability becoming the bottleneck after generation accelerates and is retained as pressure-testing evidence for the verification bandwidth trend.
SkillSpector Adds a Security Gate for Agent Skills
NVIDIA's scanner treats installable agent instructions as executable supply-chain artifacts that require inspection before use.
Walkinglabs / Learn Harness Engineering: Verification Bandwidth
The archive captures Walkinglabs / Learn Harness Engineering as a dated public record from Walkinglabs / Learn Harness Engineering. It documents evaluation, review, and observability becoming the bottleneck after generation accelerates and is retained as pressure-testing evidence for the verification bandwidth trend.
Claude for Legal Packages a Legal Team as Plugins
Anthropic's legal repository expresses domain knowledge, procedures, and tools as installable modules for repeatable professional work.
Claude: Goal-Scoped Agent Loops
The archive captures YouTube source as a dated public record from YouTube source. It documents long-running work gaining explicit goals, state, stopping rules, and recovery and is retained as supporting evidence for the goal-scoped agent loops trend.
SKILL.md Acts as a Behavior Loader, Not a Better Prompt
Skills package procedures, tools, and progressive context so an agent can load behavior only when the task requires it.
Claude: Goal-Scoped Agent Loops
The archive captures X source as a dated public record from X source. It documents long-running work gaining explicit goals, state, stopping rules, and recovery and is retained as pressure-testing evidence for the goal-scoped agent loops trend.
Cursor Ships Its Development Procedures as Agent Skills
Cursor Team Kit packages verification, CI repair, review preparation, and code cleanup as reusable agent behavior.
GitHub / NousResearch/hermes-agent: Generative Media Infrastructure
The archive captures GitHub / NousResearch/hermes-agent as a dated public record from GitHub / NousResearch/hermes-agent. It documents image, video, audio, and multimodal generation becoming application infrastructure and is retained as branch-opening evidence for the generative media infrastructure trend.