The Agent Skills Directory

[SAFE]: The skill's primary functionality is implemented through local Python scripts that handle text processing and local JSON storage. No suspicious commands, obfuscated code, or unauthorized network operations were detected.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data (such as participant names, agendas, and meeting titles) to populate markdown templates, creating a standard surface for indirect injection if malicious text is provided as input. Ingestion points: Command-line arguments in scripts/meeting_brief_generator.py and scripts/action_item_tracker.py. Boundary markers: None; the generated markdown output does not use delimiters to isolate user-provided strings. Capability inventory: The skill is limited to writing local files (action item database and briefing documents) and does not possess network or administrative capabilities. Sanitization: No explicit sanitization or filtering is performed on the input strings before they are incorporated into the briefing markdown.

legal-meeting-briefing