Agent Skills
Also called skills, SKILL.md, progressive disclosure.
Agent skills are packaged folders of instructions, scripts, and reference files, each anchored by a SKILL.md file with a name and description, that an agent loads only when a task calls for them.
Description
Only each skill's short description sits in the agent's context at startup. The full instructions are read when relevant, and supporting files are opened only if needed, a pattern called progressive disclosure. This lets an agent carry many specialized procedures without filling its context window, and lets organizations package their own workflows for reuse.
Sources
- Anthropic (2025). Equipping agents for the real world with Agent Skills.
Cite this entry
Protologue. (2026). Agent Skills. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0148). https://protologue.com/t/agent-skills/
BibTeX
@misc{protologue_agent_skills,
title = {Agent Skills},
author = {{Protologue}},
year = {2026},
howpublished = {Protologue: A Taxonomy of Prompting and LLM Techniques, v1.4.0},
note = {Entry PTL-0148},
url = {https://protologue.com/t/agent-skills/}
}New terms, as they're coined
One email a week with the newest techniques, each defined and sourced like this entry.
Delivered through Beehiiv. Unsubscribe from any issue.