extension-oql
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the "caffeineai-oql" package via the "mops" package manager, which is a resource provided by the vendor.
- [COMMAND_EXECUTION]: Instructions are provided for the user to execute "mops" commands to install the necessary Motoko libraries.
- [PROMPT_INJECTION]: The skill establishes a path for the agent to ingest canister data, creating a surface for indirect prompt injection.
- Ingestion points: Data retrieved from the backend canister through the "execute()" method.
- Boundary markers: Not explicitly defined in the skill documentation; the implementation relies on the agent's internal handling of structured query results.
- Capability inventory: Includes the "schema()" discovery tool and "execute()" data retrieval tool.
- Sanitization: The skill mitigates risks through comprehensive backend authorization levels (.public_, .controllerOnly, .scopedPerUser, .controllerOrScoped) and row-level security guards (.ownedBy, .ownedByWith).
Audit Metadata