tend
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use local CLI tools (
allium check,allium analyse) for verifying the syntax and integrity of specification files. These tools are project-specific and used for validation. - [EXTERNAL_DOWNLOADS]: Contains informational references to the
alliumCLI availability on established package registries like Homebrew and crates.io for optional automated checking. - [PROMPT_INJECTION]: The skill ingests user requirements to modify specifications, creating a surface for indirect prompt injection.
- Ingestion points: User requests for system behavior (SKILL.md)
- Boundary markers: None
- Capability inventory: Local command execution via the
alliumCLI (SKILL.md) - Sanitization: None specified
Audit Metadata