@agent-sh/harness-skill
v0.2.0
Published
Skill tool for AI agent harnesses — authored SKILL.md activation with progressive disclosure, permission-gated activation, trust-gated project skills, and pluggable SkillRegistry
Downloads
96
Maintainers
Readme
@agent-sh/harness-skill
Authored Agent Skills activation with progressive disclosure, permission-gated activation, and trust-gated project skills.
Part of the @agent-sh/harness-* monorepo — see the top-level README for architectural context and the full tool surface.
Install
npm install @agent-sh/harness-skillRequires Node ≥ 20.
Usage
import { skill, FilesystemSkillRegistry } from "@agent-sh/harness-skill";
const session = {
cwd: process.cwd(),
permissions: { roots: [process.cwd()], sensitivePatterns: [], unsafeAllowSkillWithoutHook: true },
registry: new FilesystemSkillRegistry([".skills"]),
trust: { trustedRoots: [process.cwd()] },
activated: new Set<string>(),
};
const r = await skill({ name: "api-conventions" }, session);Contract
The full contract — input shape, output discriminated-union, error codes, permission model, and acceptance tests — lives in agent-knowledge/design/skill.md. Changes to this package must stay in sync with that spec.
License
MIT © Avi Fenesh
