stackbone-coder
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via a guided interview process to generate code files and execute CLI commands. 1. Ingestion points: User responses collected during the scaffolding interview in SKILL.md and reference files. 2. Boundary markers: Absent; user input is interpolated into generated code blocks without explicit delimiters. 3. Capability inventory: File system write operations for generating TypeScript files and shell command execution via the stackbone-cli skill. 4. Sanitization: No explicit sanitization or validation of user-provided prompt strings is described.
- [DYNAMIC_EXECUTION]: The skill is designed to generate TypeScript source code (e.g., index.ts, schema.ts) based on user requirements, which are then executed or emulated using the stackbone dev tool.
- [EXTERNAL_DOWNLOADS]: The skill identifies and references external developer resources, specifically documentation for LangChain's deepagents library and the Workflow SDK, to assist in the configuration of the generated projects.
Audit Metadata