init
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior or suspicious code patterns were identified. The skill correctly uses platform-provided tool restrictions to ensure the agent only executes intended commands and does not have access to unauthorized system utilities.
- [INDIRECT_PROMPT_INJECTION]: The skill uses user input to determine command flags for the git-agent CLI. This represents an intended functional surface for the tool and is secured by scoping the bash tool access to the git-agent binary.
- Ingestion points: User-provided input via the $ARGUMENTS variable in SKILL.md
- Boundary markers: None present in the instruction set
- Capability inventory: Bash command execution restricted by the allowed-tools configuration to the git-agent binary
- Sanitization: None detected in the skill instructions; relies on tool-scoping for safety
Audit Metadata