post-formatter
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions mandate the use of shell utilities (
wc -l) and a Python script (skills/shared/scripts/wordcount.py) to perform deterministic validation of the output's length and line counts during the self-check step. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied themes and notes for formatting. Although it lacks explicit boundary markers for these inputs, the potential for exploitation is minimal as the agent's task is limited to structural transformation using marketing frameworks.
- [METADATA_POISONING]: Metadata in
EASEL-META.mdprovides clear attribution to open-source sources and contributors without deceptive intent or hidden instructions.
Audit Metadata