pdfcraft-editor
Warn
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to manually execute a shell command to start a server if the tools are not registered. This bypasses tool abstractions and allows the agent to run arbitrary scripts within the workspace.
- Evidence: "可在得到插件实际目录后用项目终端运行
node ui/dist/adapter.mjs --serve" - [INDIRECT_PROMPT_INJECTION]: The skill processes external PDF documents, which serves as an attack surface for instructions embedded within the files to influence the agent's behavior.
- Ingestion points: Processes document content via
doc_openand related document handling tools inSKILL.md. - Boundary markers: No specific boundary markers or instructions to ignore embedded content are defined.
- Capability inventory: The agent can write files (
doc_save), render content (page_render), and execute terminal commands (node). - Sanitization: No sanitization or validation methods are mentioned for the content of the documents being processed.
Audit Metadata