lov-event-curator
Warn
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill requires running 'npx lovstudio' and 'uvx lovstudio-skill-helper', which download and execute code from external package registries (NPM and PyPI). While these tools are associated with the author, they represent a dependency on external remote code for the skill to function.
- [COMMAND_EXECUTION]: The setup process instructs the user or agent to execute shell commands ('npx', 'uvx') to install and decrypt the skill payload.
- [OBFUSCATION]: The skill's primary instructions are stored in an encrypted binary file ('SKILL.md.enc') and managed by an encrypted manifest ('MANIFEST.enc.json'). This encryption hides the skill's internal logic and prevents standard safety auditing.
- [INDIRECT_PROMPT_INJECTION]: The skill processes 'guest background material' as input, which is a vector for indirect prompt injection if the source data contains malicious instructions designed to manipulate the agent.
- Ingestion points: Processes external guest background material at runtime (SKILL.md).
- Boundary markers: No explicit markers or instructions to ignore embedded commands are present in the readable files.
- Capability inventory: Setup involves shell command execution; however, the full capabilities of the decrypted payload are unknown due to encryption.
- Sanitization: No sanitization or filtering logic is visible in the provided distribution files.
Audit Metadata