ontology-manager
Fail
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/validate_ontology.shis vulnerable to command injection. It extracts regex patterns directly from a user-supplied YAML file and interpolates them into a shell command string that is executed bypython3 -c. A malicious user can provide a pattern that breaks out of the Python string context to execute arbitrary Python code and system commands. To mitigate this, input variables should be passed to the Python interpreter as command-line arguments (e.g., usingsys.argv) rather than being interpolated into the script string. - [REMOTE_CODE_EXECUTION]: The vulnerability in
scripts/validate_ontology.shallows for remote code execution if the agent is tasked with validating a malicious ontology file provided by an external or untrusted source. - [EXTERNAL_DOWNLOADS]: The
SKILL.mdinstructions recommend installing common packages and utilities (jsonschema,pyyaml,yq,jq) from well-known repositories using standard tools likepipandbrew.
Recommendations
- AI detected serious security threats
Audit Metadata