matter-update

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's functionality is limited to reading and writing local files related to a legal workflow. No evidence of malicious intent or dangerous patterns was found.
  • [COMMAND_EXECUTION]: There are no shell commands or subprocess calls. The skill purely uses natural language instructions for the agent to manipulate files within the specified paths.
  • [DATA_EXPOSURE]: The skill accesses files located in ~/.claude/plugins/config/claude-for-legal/. These are application-specific configuration files rather than sensitive system-level files like credentials or SSH keys.
  • [PROMPT_INJECTION]: The skill processes existing case files (history.md, _log.yaml) to perform updates. While this presents a minor theoretical surface for indirect prompt injection if those files were compromised, the skill mandates a user confirmation step ("This is what I will append... Is it okay to write?") before any persistent changes are made, providing a human-in-the-loop safeguard.
  • [EXTERNAL_DOWNLOADS]: No external network requests, downloads, or remote script executions are present.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 07:06 AM
Security Audit — agent-trust-hub — matter-update