sparkscan-capacitor
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest user input (barcode symbology requirements and target file paths) and write code directly into the user's project files. While this creates a surface for indirect prompt injection if the user's existing codebase contained malicious instructions, the skill includes guardrails directing the agent to verify all APIs against provided reference documents rather than relying on unverified data.
- Ingestion points: Processes user-provided requirements and file paths in
references/integration.mdandreferences/migration.md. - Boundary markers:
SKILL.mdincludes explicit instructions to 'Always verify APIs against the references provided in this skill before writing or suggesting code.' - Capability inventory: The skill allows writing JavaScript/TypeScript code to local files (
app.js,main.ts) as described inreferences/integration.md. - Sanitization: The instructions focus on structured SDK implementation, reducing the risk of arbitrary code execution from user-supplied strings.
Audit Metadata