little-steve-agent-guard - Self-evolving security system for agent skills
Wraps skill commands with risk assessment, audit logging, tiered approval, and continuous rule learning.
Tags
Updated: 2026-05-09Capabilities
Typical Inputs
Typical Outputs
What this skill does
- execute skill command
- preview script execution
- check risk level
- view audit statistics
- generate security report
- list rules
- add rule
- promote rule
- demote rule
- test rule
- test all rules
- check capability consistency
- read target scripts
Inputs
- skill script path
- command arguments
- rule name
- rule pattern
- risk level
- skill directory
- confirmation response
- audit log
Outputs
- audit log file
- security report
- rule list
- execution warning
- capability difference report
- risk assessment result
- test result
Requirements
- jq