session-recovery

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFE
Full Analysis
  • Local Data Access: The skill uses the sql tool to query a local database (session_store) to retrieve session metadata and history. This is a standard and necessary function for session recovery, limited to the specific database intended for this purpose.
  • Indirect Prompt Injection Surface: The skill processes data from past conversation turns (user_message), which technically presents a surface for indirect prompt injection if previous sessions contained malicious content.
  • Ingestion points: The turns and sessions tables in the session_store database, accessed via the sql tool.
  • Boundary markers: None are explicitly used in the provided SQL search patterns.
  • Capability inventory: The skill facilitates the copilot --resume command execution.
  • Sanitization: There is no specific sanitization of the retrieved text before it is presented back to the agent. This configuration is inherent to the task of workflow recovery and is generally managed by the agent's internal safety controls.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 10:33 AM
Security Audit — agent-trust-hub — session-recovery