extension-querying-oql

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the icp CLI tool to interact with the project's backend canister. Examples include icp canister call backend schema '()' --query and icp canister call backend execute '("<json-query>")' --query.- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and process data returned from a remote canister, which represents a surface for indirect prompt injection if the retrieved content contains malicious instructions.
  • Ingestion points: Data returned from icp canister call backend schema and execute methods as documented in SKILL.md.
  • Boundary markers: Absent; there are no instructions to use delimiters or ignore instructions embedded within the canister's output.
  • Capability inventory: The agent can execute shell commands via the icp CLI and perform autonomous query construction based on the schema.
  • Sanitization: Absent; the instructions do not specify any validation or sanitization of the data retrieved from the canister before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 09:39 PM
Security Audit — agent-trust-hub — extension-querying-oql