The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration were detected. The skill's behavior is consistent with its stated purpose of generating content outlines.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests and processes content from external websites.
Ingestion points: Data enters the context via web_fetch on the user-provided sitemap_url and through dataforseo:on_page_content_parsing in SKILL.md (Step 3).
Boundary markers: No explicit XML/Markdown delimiters are instructed for the ingested content, though the data is summarized into a domain_context variable.
Capability inventory: The skill can execute a local Python script (scripts/generate-brief.py) and create files in the /mnt/user-data/outputs/ directory.
Sanitization: The skill mitigates risks through a mandatory 12-point quality check (Step 7) and strict structural rules in references/section-rules.md that prevent the agent from including conclusions, invented numbers, or long prose influenced by the source data.

brief-outline-generator