journey
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [External Downloads] (MEDIUM): The documentation suggests installation using
npx clawhub@latestand a specific GitHub repositorybarneyjm/camino-skills. Neither the tool nor the repository are from trusted organizations or listed in the security guidelines, presenting a risk of unverifiable code execution during installation. - [Data Exfiltration] (LOW): The skill's primary function involves sending a user-provided
CAMINO_API_KEYand journey waypoints tohttps://api.getcamino.ai. While necessary for the service, users should be aware of this data transfer to a non-whitelisted domain. - [Indirect Prompt Injection] (LOW): The skill processes external waypoint data and returns a 'summary' from the API. This creates a surface for Indirect Prompt Injection if the agent relies on this summary for subsequent decision-making. * Ingestion points: waypoints array in
scripts/journey.sh. * Boundary markers: None identified in the script or instructions. * Capability inventory: Subprocess calls tocurlandjqfor network communication and data parsing. * Sanitization: Input is validated as JSON viajq, but internal string content is not sanitized.
Audit Metadata