jason-corpus-search

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a bash wrapper script (search.sh) to execute a local command-line interface tool (cli storage search) located at /Users/jcohen/git/rt. The script correctly handles agent-provided arguments using quoted parameters, following standard safe practices for shell wrappers.- [PROMPT_INJECTION]: The skill provides the agent with the ability to read and process external markdown content, which creates a surface for indirect prompt injection.
  • Ingestion points: Markdown files read from the /Users/jcohen/Obsidian/ directory structure via the Read tool.
  • Boundary markers: No specific delimiters or safety instructions are defined to separate ingested data from agent instructions.
  • Capability inventory: The skill allows Bash (local script), Read, Grep, and Glob operations.
  • Sanitization: The skill does not perform sanitization or validation of the content within the markdown files before they are presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 05:48 PM
Security Audit — agent-trust-hub — jason-corpus-search