web-pwa-offline-first

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for handling user-authored records that create a surface for indirect prompt injection when used by an AI agent to process or sync untrusted data.\n
  • Ingestion points: Untrusted data enters the context through repository save/update methods and conflict resolution interfaces (e.g., TodoRepository in examples/core.md and ConflictResolver in examples/sync.md).\n
  • Boundary markers: The templates lack explicit markers or instructions to delimit user data from system instructions during processing.\n
  • Capability inventory: The skill implements network operations using the fetch API to synchronize local changes to remote servers (e.g., SyncQueue in examples/core.md).\n
  • Sanitization: There is no evidence of data sanitization or validation in the provided code snippets prior to storage or network transmission.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 03:07 PM
Security Audit — agent-trust-hub — web-pwa-offline-first