mcloud-environments
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes JSON output from the mcloud CLI, which represents an ingestion point for potentially untrusted data if environment metadata is controlled by an external party.
- Ingestion points: The agent reads and processes output from
mcloud environments list --jsonandmcloud environments get --jsonas described in SKILL.md. - Boundary markers: Absent; there are no specific instructions for the agent to distinguish between data and instructions within the tool output.
- Capability inventory: The skill's capabilities are strictly limited by the
allowed-toolsfrontmatter tomcloud environments,mcloud use, andjqcommands, preventing arbitrary shell execution. - Sanitization: Absent; the skill relies on the agent's interpretation of structured JSON without explicit sanitization of text fields. However, the restricted tool scope effectively mitigates the risk of command injection or privilege escalation.
Audit Metadata