v-skill-creator
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs Git operations (status, add, commit, push) and utilizes the GitHub CLI (gh) to manage repository visibility and publish new skill content to the MamounHisham1/vstack repository.\n- [EXTERNAL_DOWNLOADS]: The workflow uses the 'npx skills add' command to fetch and install skills from the vendor's remote repository, involving the download and execution of instructions from an external source.\n- [INDIRECT_PROMPT_INJECTION]: The skill scaffolds command templates that interpolate user-provided arguments directly into a prompt, which could allow malicious input to override agent behavior.\n
- Ingestion points: The $ARGUMENTS variable within the generated .opencode/commands/.md files.\n
- Boundary markers: None; user-supplied data is directly concatenated to the agent's instructions.\n
- Capability inventory: Scaffolding files, Git repository management, and running npx-based installation tools.\n
- Sanitization: While static file content is sanitized for credentials before commit, there is no mechanism to sanitize or isolate the runtime user input in the resulting commands.
Audit Metadata