The Agent Skills Directory

[SAFE]: No malicious code, obfuscation techniques, or unauthorized resource access patterns were detected within the skill's instructions or schema definitions.
[DATA_EXFILTRATION]: The skill mandates strict data handling protocols, explicitly stating that secrets, cookies, private IDs, and account-specific values must not be recorded. It also prohibits documenting methods to bypass CAPTCHAs, WAFs, or other access controls.
[COMMAND_EXECUTION]: The skill uses the webcmd browser tool suite to analyze site state, which is appropriate for its purpose and restricted to the permitted tool scope.
[PROMPT_INJECTION]: While the skill interacts with untrusted external website content, it mitigates indirect injection risks through structured output requirements and sanitization instructions.
Ingestion points: External site data retrieved via webcmd browser (SKILL.md).
Boundary markers: Requirement for structured Markdown/YAML schemas for all sitemap files (references/sitemap-schema.md).
Capability inventory: Tools for reading, writing, and executing webcmd bash commands (SKILL.md).
Sanitization: Explicit instructions in the 'Red Lines' section to filter out private data and ignore unverified paths (SKILL.md).

webcmd-sitemap-author