deploy
Warn
Audited by Gen Agent Trust Hub on Jul 26, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses dynamic context injection (exclamation mark followed by backticks) to execute
dateandkubectl config current-contextautomatically when the skill is loaded. While these commands are common in deployment workflows, their silent execution at load time discloses environment and infrastructure details to the agent context without explicit user consent.\n- [PROMPT_INJECTION]: The skill ingests untrusted user input for deployment targets and scope, which is subsequently processed to generate documentation via theobsidian_append_contenttool. It lacks specific instructions for the agent to treat this data as passive content, creating a risk of indirect prompt injection.\n - Ingestion points: User-provided deployment target, scope, and strategy fields.\n
- Boundary markers: Not present; user input is not separated from the main prompt instructions.\n
- Capability inventory: The agent has the ability to write to the local filesystem using Obsidian tools.\n
- Sanitization: The skill does not define any validation or escaping mechanisms for the external input data.
Audit Metadata