skills/wojons/skills/ai-collab-dev/Gen Agent Trust Hub

ai-collab-dev

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected during the audit. The skill consists of helper scripts for local file management and project scaffolding.
  • [COMMAND_EXECUTION]: The skill provides bash scripts (chunk-conversation.sh, setup-ralph-project.sh, validate-todo.sh) that perform file operations, directory creation, and text processing. These actions are aligned with the skill's stated purpose of managing development workflows.
  • [DYNAMIC_EXECUTION]: The setup-ralph-project.sh script dynamically creates a project environment, including a Python runner (ralph-loop.py) with execution permissions. It also attempts to integrate with the ralph-wiggum-loop skill by copying existing templates from local paths if available.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external text files (AI conversation logs) through chunk-conversation.sh.
  • Ingestion points: scripts/chunk-conversation.sh reads content from a user-specified input file for splitting.
  • Boundary markers: Messages are delineated by standard prefixes such as Human:, User:, or Me:.
  • Capability inventory: The skill is primarily a text-processing and scaffolding tool; command execution is limited to the provided scripts and the generated project runner which is intended for local use.
  • Sanitization: No sanitization is performed on the input conversation text as it is intended for manual transfer by the developer.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 10:02 AM