skills/johngrimes/mojoflow/spec/Gen Agent Trust Hub

spec

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a specific git command (git rev-parse --path-format=absolute --git-common-dir) to identify the main repository root. This is used to ensure specification files are stored consistently within the project structure.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting untrusted data from multiple sources:
  • User Input: It processes feature descriptions provided via $ARGUMENTS.
  • Project Files: It reads project-level configuration files (specifically CLAUDE.md) to perform a "Constitution Check" and detects the existing codebase stack and conventions.
  • Evidence Chain: Ingestion points include the user prompt and CLAUDE.md in the repository root. The skill has capabilities to write files and execute shell commands. There are no explicit instructions for sanitizing these inputs before interpolation into the generation process.
  • [SAFE]: The skill operates locally using provided templates and does not initiate network connections, external downloads, or remote code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 11:20 AM
Security Audit — agent-trust-hub — spec