lov-better-css
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PERSISTENCE]: The skill includes a utility script (
scripts/profile_store.py) designed to maintain persistent user and brand identity data. It stores information in a JSON profile located at~/.config/agent-skills/profile.jsonor~/.lovstudio/skills/profile.json, utilizing atomic writes and permission controls to manage long-term state across sessions. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a potential attack surface for indirect prompt injection as it processes external project files.
- Ingestion points: Reads stylesheets and component files during the analysis phase (
SKILL.md). - Boundary markers: No explicit boundary markers or instructions to ignore embedded content are provided for the CSS/component ingestion.
- Capability inventory: The skill has the capability to modify project files, run build commands, and perform type checks (
SKILL.md). - Sanitization: No sanitization or filtering of the ingested CSS content is mentioned before it is processed by the agent.
- [COMMAND_EXECUTION]: The execution workflow involves running local project build tools and type checkers to validate refactoring results, which is consistent with its stated purpose of improving project styles.
Audit Metadata