The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local scripts and system tools (Git, PHP Artisan) to inspect project structure, detect frameworks, and verify environment safety. For example, it uses php artisan migrate:status to check the state of database migrations.
[EXTERNAL_DOWNLOADS]: Installs additional specialized skills from the author's official repository using npx skills add https://github.com/promovaweb/specsfy. This is a standard extension mechanism for the Specsfy framework to provide modular features based on the detected project stack.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data (e.g., files in specs/, PROJECT.md) to build the agent's context and generate user questions. This represents a potential surface for instructions embedded in project documentation to influence agent behavior.
Ingestion points: Reads project-specific files including PROJECT.md, .specsfy/USER-PROFILE.md, and all files within the specs/ directory.
Boundary markers: Utilizes specific markdown delimiters (e.g., ) to manage blocks, though it relies on general LLM comprehension for non-managed content.
Capability inventory: Can execute shell commands for project inspection and install additional skill modules via npx.
Sanitization: Employs template-based substitution and structure-based parsing of configuration files to maintain document integrity.

specsfy-setup