The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides commands in SKILL.md to download an agent installer from Akamai's official Guardicore management domain (management.guardicore.com). This is followed by local execution to enroll workloads into the segmentation platform.
[COMMAND_EXECUTION]: The scripts/agent.py script utilizes the subprocess module to execute curl for interacting with the Guardicore Centra API. It passes API tokens and JSON payloads as command arguments. The script specifically uses the -k (insecure) flag in curl, which disables SSL certificate verification—a common configuration in internal environments that nonetheless carries risk of man-in-the-middle attacks.
[PRIVILEGE_ESCALATION]: The instructions require the use of sudo for installing the Linux agent and configure Kubernetes DaemonSets with the privileged: true security context. These elevated permissions are functionally necessary for microsegmentation agents to monitor host network traffic and enforce granular policies.
[INDIRECT_PROMPT_INJECTION]: The audit script in scripts/agent.py processes data from network flow logs and security policies, presenting an attack surface for indirect prompt injection if external data is malicious.
Ingestion points: Network connection telemetry and policy definitions retrieved from the Guardicore Centra API.
Boundary markers: None identified in the script logic or prompts.
Capability inventory: Shell execution via subprocess.run and file output using json.dump.
Sanitization: The script uses standard JSON parsing to handle API responses but does not implement validation or filtering of specific data fields, such as workload labels or process names, before including them in the generated audit report.

implementing-microsegmentation-with-guardicore