paperclip-create-plugin
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using
pnpmandnode. This includes building a local package (@paperclipai/create-paperclip-plugin) and running the resulting JavaScript artifact (dist/index.js). While common in development environments, this grants the agent the ability to execute code within the local file system. - [INDIRECT_PROMPT_INJECTION]: The skill contains a command injection surface in the preferred workflow section. User-provided values for
<npm-package-name>and<target-dir>are interpolated directly into shell commands (node ... <npm-package-name> --output <target-dir>).- Ingestion points: File:
SKILL.md. User input for package names and directory paths. - Boundary markers: Absent. There are no instructions to validate or sanitize these inputs before use.
- Capability inventory:
pnpm,node. - Sanitization: Absent. No escaping or validation steps are defined.
- Ingestion points: File:
Audit Metadata