skills/jkitchin/scimax_vscode/scimax/Gen Agent Trust Hub

scimax

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides an executive assistant interface for the scimax-vscode extension and scimax CLI. All identified tool interactions, including command execution for the scimax CLI and file system operations, are consistent with the objective of managing Org-mode notes and workflows.
  • [PROMPT_INJECTION]: The skill ingests and processes user-controlled data from Org-mode files and local learning logs. While this creates a potential surface for indirect prompt injection, it is a necessary functional requirement for a note-taking assistant.
  • Ingestion points: Reads user-defined .org files and the skill's own learnings.md file using the Read tool.
  • Boundary markers: The prompt does not utilize explicit delimiters or 'ignore' instructions for the content being summarized or extracted from notes.
  • Capability inventory: The skill is permitted to run scimax CLI subcommands, use code --goto for file navigation, and write updates to the local learnings.md file.
  • Sanitization: Note content is handled as plain text without filtering for embedded agent instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 08:20 PM
Security Audit — agent-trust-hub — scimax