publish-project-to-github

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust security posture by requiring a mandatory audit via a dedicated script (scripts/audit_public_project.sh) before packaging or publishing any code. This script specifically searches for API keys, private keys, environment variables (.env), and personal filesystem paths.
  • [DATA_EXFILTRATION]: While the skill's purpose is to upload data to a remote server (GitHub), it explicitly requires user authorization for public visibility and uses standard, authenticated tools (git and the GitHub CLI). It targets a well-known service (GitHub) and includes instructions to verify state and avoid overwriting existing data without permission.
  • [COMMAND_EXECUTION]: The skill uses local shell commands for legitimate administrative tasks, such as initializing Git repositories, creating remote repositories via the GitHub CLI, and running a local development server for verification. These commands are constrained to the intended project scope.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local project files. It mitigates injection risks through multiple defensive layers:
  • Ingestion points: The project directory and files are inspected during the Resolve and Audit steps.
  • Boundary markers: A strict README template is used to structure generated documentation.
  • Capability inventory: Capabilities are limited to standard Git/GitHub operations and a local Python-based preview server.
  • Sanitization: The audit script acts as a validator, blocking the process if suspicious content or credentials are detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 12:00 PM
Security Audit — agent-trust-hub — publish-project-to-github