The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses standard identity and role-definition framing without any attempts to override safety filters, disregard prior instructions, or extract system prompts.
[DATA_EXFILTRATION]: No sensitive file paths, credential harvesting patterns, or unauthorized network exfiltration attempts were identified. Network references are limited to official documentation, the author's public GitHub repository for installation, and well-known service APIs (GitHub API).
[REMOTE_CODE_EXECUTION]: No patterns of downloading and executing arbitrary remote scripts (e.g., curl|bash) were found. Command snippets provided in the technical guides are for documentation and administrative use (PowerCLI, Tanzu CLI) and do not perform malicious actions.
[COMMAND_EXECUTION]: Shell and PowerShell snippets are strictly for administrative documentation (e.g., managing vSphere clusters, configuring NSX rules). No automated or silent execution of dangerous commands is present.
[OBFUSCATION]: The content is entirely human-readable Markdown. No Base64, hex encoding, zero-width characters, or homoglyph-based obfuscation techniques were detected in the instructions or URLs.
[CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or private keys were found. Placeholders like '' are used appropriately for documentation.
[DYNAMIC_EXECUTION]: The skill does not generate or execute code at runtime using eval(), exec(), or similar functions. All technical examples are static documentation.
[PRIVILEGE_ESCALATION]: No commands for unauthorized privilege escalation (e.g., sudo, chmod 777) or persistence mechanisms were detected.
[INDIRECT_PROMPT_INJECTION]: The skill is a static knowledge base and does not provide an attack surface for ingesting untrusted external data into the agent's prompt context.
[METADATA_POISONING]: Metadata fields are consistent with the skill's technical purpose and author identity.

vmware-engineer