TP
anthropic / trusted-skill-import
Hosted SKILL.mdTrusted provider provenance example
docs.claude.com/agents-and-tools/agent-skills
Review a skill obtained from a trusted external provider before admitting it to a local AutoVault.
InstallCLI + MCP
$autovault add https://autovault.dev/skills/trusted-skill-import/SKILL.md --source url --sync-profiles
>add_skill({ source: "url", identifier: "https://autovault.dev/skills/trusted-skill-import/SKILL.md" })
Choose CLI for a shell install or MCP for an agent tool call.
Example type
skill
hosted SKILL.md
Declared agents
2
from frontmatter
Gate stages
5
covered by tests
Permission rows
3
declared metadata
Source
Anthropic
Trusted provider provenance example
---
name: trusted-skill-import
version: 0.1.0
description: Review a skill obtained from a trusted external provider before admitting it to a local AutoVault.
source.provider: Anthropic
capabilities.filesystem: readonly
tools_required: Read
---
trusted-skill-import
Shows how AutoVault can represent a trusted external provider source while still requiring local review.
Keeps install behavior explicit: users copy the MCP add_skill call or stage the file locally before running the admission gate.
When to use this skill
- A user brings a skill from a trusted external provider.
- The source URL and local copy need provenance review before admission.
- A team wants to show third-party trust context without operating a marketplace.
Install
CLI: autovault add https://autovault.dev/skills/trusted-skill-import/SKILL.md --source url --sync-profiles
MCP: add_skill({ source: "url", identifier: "https://autovault.dev/skills/trusted-skill-import/SKILL.md" })
Provenance
Models a trusted-provider review flow using Anthropic's public skills guidance without making the catalog a marketplace.