project-claude-organizer

Pass

Audited by Gen Agent Trust Hub on Mar 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs multiple file system operations including directory creation, file writing, content appending, and file deletion. While these are gated by user confirmation, they represent a significant capability for modifying the local project environment.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8).
  • Ingestion points: It reads content from multiple user-controlled files within the .claude/ directory, specifically from subfolders like commands/, system/, and sops/, as well as files like project.md and readme.md (Step 3b, Step 5.7).
  • Boundary markers: Appended content is preceded by a markdown comment (e.g., ), which provides a structural delimiter but does not prevent the AI from following instructions embedded within the appended text.
  • Capability inventory: The skill can write to ai-context/ files, create new skill files (SKILL.md) in .claude/skills/, and delete source files after migration (Step 5).
  • Sanitization: There is no significant sanitization of the content being moved; only emoji normalization is performed on section headings to facilitate routing logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 22, 2026, 09:45 AM