The Agent Skills Directory

[SAFE]: The skill implements strong security-oriented patterns as part of its core architecture. Specifically, pattern STR-03 (Workspace Isolation) explicitly instructs generated agents to work only within a designated project directory, prohibiting access to external system files or other projects.
[SAFE]: The skill addresses potential data integrity issues through patterns QUA-03 (Verifiable Data Lineage) and BHV-05 (Grounded Web Research). These require agents to fetch and preserve original source content verbatim and use independent verification steps to prevent hallucinations or the ingestion of distorted data.
[SAFE]: The provided utility script templates, such as export.sh and assemble.sh, use standard local tools (cat, pandoc, xelatex) for document processing. They do not involve remote code execution, obfuscation, or unauthorized network operations.
[SAFE]: Tool bindings mentioned in BHV-06 (e.g., serper, crawl4ai, oxylabs) refer to well-known search and web-fetching services commonly used in AI development environments and do not represent a security risk within the context of the skill's research purpose.

pomasa