skills/motiful/ctx/ctx-report/Gen Agent Trust Hub

ctx-report

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill distills human comments into the project source of truth, creating a vulnerability surface where instructions in comments could be followed by the agent. * Ingestion points: Human comments on HTML reports. * Boundary markers: None specified to isolate user input. * Capability inventory: Filesystem writing via ctx-merge. * Sanitization: No sanitization rules defined for comment content.
  • [EXTERNAL_DOWNLOADS]: The skill generates reports that fetch the Mermaid.js library and pan-zoom tools from the well-known jsdelivr.net CDN.
  • [DATA_EXFILTRATION]: The skill requires absolute filesystem paths (e.g., /Users/...) in reports, which exposes host system directory structures to the output.
  • [DYNAMIC_EXECUTION]: The skill generates HTML script modules to load external libraries, involving the generation of code meant for client-side execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 06:35 AM
Security Audit — agent-trust-hub — ctx-report