add-remote-skill
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to execute a local TypeScript utility (
bin/add-remote-skill.ts) which clones content from remote GitHub repositories. This provides a mechanism for bringing external, untrusted code into the local agent environment. - [INDIRECT_PROMPT_INJECTION]: The skill ingests
SKILL.mdfiles from external repositories and integrates them into the marketplace. This represents an attack surface for indirect prompt injection where malicious instructions in a remote repository could influence the agent's behavior once the new skill is loaded. - Ingestion points: Remote GitHub repositories and their constituent
SKILL.mdfiles. - Boundary markers: No explicit delimiters or instructions to ignore embedded prompts are mentioned during the ingestion process.
- Capability inventory: The skill uses
git sparse-checkoutto pull remote files and performs file writes to the localskills/directory. - Sanitization: No sanitization, validation, or safety filtering of the downloaded markdown content is described in the workflow.
Audit Metadata