The Agent Skills Directory

[SAFE]: The skill operates as a routing and compliance layer, focusing on identifying task types and applying relevant legal and stylistic overlays. No evidence of malicious intent, unauthorized command execution, or data exfiltration was found.
[INDIRECT_PROMPT_INJECTION]: The skill processes external prospect data and user-provided call intel, which constitutes an ingestion surface for potential indirect instructions.
Ingestion points: User-provided prospect research, discovery call notes, and company profiles ingested via SKILL.md and routed to specialized sub-skills.
Boundary markers: Implements a 'MANDATORY OUTPUT HEADER' and the 'Five Laws of Outreach' compliance check to enforce structure and validation on all outputs.
Capability inventory: The routing skill contains no direct capabilities for file writing, network requests, or system command execution; it functions strictly as a logic router to other skill files.
Sanitization: The instructions explicitly forbid the fabrication of prospect data and mandate human review for all generated outreach content.
[METADATA_POISONING]: Metadata fields, including the descriptive routing terms in SKILL.md, are used appropriately for task identification and do not contain deceptive instructions or injection attempts.

sales-marketing-global-router