barry-coach

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends using the skills utility via npx to download and install the InterpreterBarry/barry-english-toolkit package. This is used to fetch methodology resources from the author's repository.
  • [COMMAND_EXECUTION]: Provides a maintenance command for the user to execute (npx -y skills add InterpreterBarry/barry-english-toolkit -g --all) to ensure the methodology wiki is correctly installed and accessible.
  • [PROMPT_INJECTION]: Identifies an indirect prompt injection surface due to data ingestion from external files.
  • Ingestion points: Reads user profile data from ~/.barry-english/profile.md and instructional content from methodology wiki files located in METHODOLOGY_ROOT/wiki/.
  • Boundary markers: Lacks explicit delimiters or boundary markers for external data within the instruction set.
  • Capability inventory: The skill can write to the local file system (fancy-vocab.md) and suggest shell commands to the user.
  • Sanitization: No formal sanitization of ingested content is described, though the skill is instructed to adapt responses using a specific persona rather than verbatim execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 02:44 PM
Security Audit — agent-trust-hub — barry-coach