The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains a mandatory workflow directive in SKILL.md that instructs the AI agent to orchestrate specialized sub-agents using the TeamCreate tool. This is a prescriptive instruction for the agent's operational logic.
[COMMAND_EXECUTION]: The templates and guides provided in the skill (e.g., creating-module.md, SKILL.md) suggest the use of standard development commands like mkdir, php artisan, and composer to manage the modular monolith structure.
[PROMPT_INJECTION]: The modular architecture involves an auto-discovery mechanism that parses module.json files from the filesystem (references/module-discovery.md), creating a surface for indirect prompt injection.
Ingestion points: module.json files located within the /FuseCore/ directory.
Boundary markers: None specified in the provided documentation or templates.
Capability inventory: The agent is expected to use file system tools, execute CLI commands, and spawn other agents.
Sanitization: The documentation does not describe any validation or sanitization for the contents of the discovery files.
[NO_CODE]: The skill is composed exclusively of documentation and templates in Markdown format, with no executable logic or scripts included.

fusecore