The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill's behavior is consistent with its stated purpose of infrastructure change analysis.- [COMMAND_EXECUTION]: The skill utilizes a local Python script (scripts/change_risk_from_topology.py) to perform structured analysis of topology data. The script is restricted to local file system operations (reading JSON files) and graph traversal algorithms using standard Python libraries.- [DATA_EXFILTRATION]: Analysis of the Python script and skill instructions shows no evidence of network operations, external requests, or unauthorized data transmission. Data remains within the local environment and the agent's context.- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it processes data from external files (topology.json and account_assets_snapshot.json) that could be manipulated. Ingestion points: Reads infrastructure metadata from topology.json and account_assets_snapshot.json via the change_risk_from_topology.py script. Boundary markers: The skill instructions do not specify delimiters for the script's output when presented to the model. Capability inventory: The skill can execute a local Python script that reads from the file system. Sanitization: The script does not perform sanitization of strings extracted from the JSON files before they are included in the output. This is considered an inherent and low-risk aspect of the skill's functionality.

byted-volcengine-topology-change-impact