context-continuity

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill functions as a context management utility operating entirely within the agent's environment. It does not access sensitive system files or bypass safety guidelines.
  • [COMMAND_EXECUTION]: The skill instructs the agent to use the 'view' command to read internal templates located at hardcoded paths (e.g., /home/claude/context-continuity/references/). This is part of the documented workflow and does not allow for arbitrary file access.
  • [DATA_EXFILTRATION]: The skill does not utilize network-connected tools or external APIs. All data processing is confined to generating text artifacts for the user to transfer manually.
  • [PROMPT_INJECTION]: The skill includes instructions for mode selection and context analysis that are consistent with its stated purpose. There are no patterns suggesting attempts to override model constraints or extract system prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 03:29 AM
Security Audit — agent-trust-hub — context-continuity