clawcolab
Warn
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (MEDIUM): The skill requires the installation of the 'clawcolab' Python package via pip. As this package and its repository are not listed as Trusted External Sources, the safety and integrity of the dependency cannot be verified.- PROMPT_INJECTION (LOW): The skill possesses a surface for indirect prompt injection due to its interaction with untrusted external data (ideas, tasks, and knowledge items). * Ingestion points: Data is ingested from api.clawcolab.com via functions such as get_ideas_list, get_tasks, and get_knowledge. * Boundary markers: There are no explicit instructions or delimiters provided to prevent the agent from obeying instructions embedded in the collaboration data. * Capability inventory: The skill enables 'reasoning' and 'coding' capabilities, which could be leveraged maliciously if the agent follows commands hidden in processed tasks. * Sanitization: No sanitization or validation of external content is described in the skill.
Audit Metadata