sd-feed-engine
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted text data such as user-provided storyboards and dialogue. This represents an indirect injection surface where malicious instructions could be embedded in the processed data. However, the skill implements a 'Gate 4' (Semantic Gate) and 'Engineering Language Cleanup' rules that explicitly strip non-visual instructions and workflow jargon, which acts as a mitigating control. The instructions also enforce a rigid 'Seven-Segment' format, limiting the agent's likelihood of following rogue instructions within the input data.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local Python scripts (
scripts/gate_check.pyandscripts/script_feed_diff.py) to validate prompt integrity. A detailed code audit of these scripts confirms they are benign text-processing utilities. They use standard libraries (re, sys, os, argparse) for regex matching and basic file I/O to check word counts, dialogue coverage, and formatting. There is no evidence of network access, credential harvesting, or unauthorized file system modifications.
Audit Metadata