iot-anomalies
Fail
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
npxto download the package@claude-flow/plugin-iot-cognitum@latestfrom the npm registry. This package does not originate from a recognized trusted organization. - [REMOTE_CODE_EXECUTION]: By using
npx, the skill executes code downloaded from a remote repository at runtime. The use of the@latestversion tag introduces a supply chain risk where a compromised version of the package could be automatically installed and executed. - [COMMAND_EXECUTION]: The skill executes shell commands using the
npxutility to run logic from the downloaded package. - [INDIRECT_PROMPT_INJECTION]: The
DEVICE_IDargument provided by the user is directly interpolated into a shell command:cognitum-iot anomalies DEVICE_ID. - Ingestion points: The
DEVICE_IDargument is accepted from user input. - Boundary markers: None present in the command template.
- Capability inventory: Shell execution via
Bashis enabled. - Sanitization: No sanitization or validation of the
DEVICE_IDis specified before it is passed to the shell. This creates a risk of command injection if a user provides a malicious string (e.g.,; rm -rf /).
Recommendations
- AI detected serious security threats
Audit Metadata