input-token-overheads
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local shell commands (
python3 -c) to execute auditing logic. This is used to calculate token lengths from local skill metadata and query internal database row counts for performance monitoring. - [DYNAMIC_EXECUTION]: The auditing procedure involves runtime execution of inline Python scripts. These scripts process directory globs, YAML configuration files, and SQLite databases stored within the user's home directory (
~/.hermes/). This activity is confined to local data gathering for optimization purposes. - [DATA_EXFILTRATION]: While the skill reads internal configuration files like
~/.hermes/config.yamland~/.hermes/mnemosyne.db, the extracted data is only used for local reporting to the agent. No network functions or external destinations were identified in the analysis.
Audit Metadata