dev-board

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The generated HTML dashboard fetches CSS and font assets from well-known services, specifically Tailwind CSS (cdn.tailwindcss.com) and Google Fonts (fonts.googleapis.com).
  • [COMMAND_EXECUTION]: The skill executes a local discovery script (scripts/discover.sh) to locate project folders and files. The implementation includes explicit instructions for the agent to use absolute paths and handle shell state carefully, which are standard safety practices.
  • [PROMPT_INJECTION]: The skill parses content from local markdown files in the .dev/ directory, which represents an indirect prompt injection surface.
  • Ingestion points: Feature PRDs (00-master-plan.md), checkpoints (checkpoint.md), and sub-PRD files.
  • Boundary markers: The board-generator sub-agent is constrained by specific extraction fields and a strict JSON output format, though raw file content is not explicitly delimited in the sub-agent prompt.
  • Capability inventory: The skill has tools for Bash, Read, and Write.
  • Sanitization: The skill performs JSON schema validation on the data returned by the sub-agent to ensure consistency before injecting it into the HTML dashboard template.
  • [SAFE]: No malicious patterns or security vulnerabilities were detected in the skill instructions or associated scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 07:17 AM