The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes npx to download and execute the @agent-native/skills and @agent-native/core packages from the npm registry. These are vendor resources provided by the author for skill installation and local plan management.
[COMMAND_EXECUTION]: The skill executes shell commands using the @agent-native/core CLI, including plan local check, plan local serve, and plan local verify. These commands are used to validate and preview MDX plans stored in the local file system.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to the processing of untrusted external content.
Ingestion points: Data enters the context via the planText argument in the create-visual-plan tool and human feedback retrieved through the get-plan-feedback tool.
Boundary markers: The skill does not provide explicit instructions or delimiters to isolate user-provided plan text or feedback from the agent's instructions.
Capability inventory: The skill has access to shell execution via npx, performs network operations to the author's hosted service at plan.agent-native.com, and writes files to the local directory in local-files mode.
Sanitization: While the documentation states that the plan renderer sanitizes HTML output, there is no mention of sanitizing or escaping the input text or feedback comments before they are used to influence the agent's plan generation.

visual-plan