The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs project-level operations including package management (pnpm/npm install, remove, update) and local verification. Specifically, it uses curl and rg (ripgrep) to check if the Lovinsp component is correctly injected into the local development server's runtime output.
[PERSISTENCE]: The skill includes a Python script (scripts/profile_store.py) used to read and write user preferences, brand information, and skill-specific records to a local JSON profile (e.g., in ~/.config/agent-skills/). This facilitates the persistence of configuration choices across different sessions.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its ingestion of project configuration files.
Ingestion points: Project files such as package.json, vite.config.ts, webpack.config.js, next.config.js, and nuxt.config.ts are read to determine the environment.
Boundary markers: None identified in the file-reading logic.
Capability inventory: The skill has the ability to install/uninstall packages and modify source code files based on the content of the ingested files.
Sanitization: The skill primarily uses keyword matching (e.g., searching for 'lovinsp') and regex to identify configuration locations rather than direct interpolation of external content into executable logic.
[EXTERNAL_DOWNLOADS]: The skill documentation describes installation via the lovstudio npx package and the installation of the lovinsp library from official package registries. These are standard developer operations for the intended functionality of the skill.

lov-integrate-lovinsp