The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform environment checks by executing node --version and sf --version to ensure the required runtime and CLI tools are available and meet the minimum version requirements.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions for the agent to guide users in installing necessary sub-skills through the command npx skills avonni/skills, which involves fetching packages from the vendor's repository.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes data from external tool outputs to build architecture plans.\n
Ingestion points: Untrusted data enters the agent context via the list_components and list_interactions tool calls to the Avonni MCP server.\n
Boundary markers: There are no explicit delimiters or specific instructions for the agent to disregard potential instructions embedded within the MCP response data.\n
Capability inventory: The skill allows for command execution (system version checks), dispatching tasks to other skills via the Skill tool, and performing read-only file existence checks.\n
Sanitization: No sanitization or validation logic is applied to the metadata retrieved from the MCP server before it is incorporated into the planning process.

avonni-architect