The Agent Skills Directory

[SAFE]: The skill provides a detailed and well-structured approach to influencer and affiliate marketing. It implements clear boundary checks, standardized output formats, and relies on established marketing tools and metrics. There are no indications of prompt injection, data exfiltration, or unauthorized command execution.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it is designed to analyze external creator content, including social media comments and historical posts, for vetting and diagnostic purposes.
Ingestion points: External data enters the agent's context through the analysis of influencer social media profiles, content history, and audience comments as described in influencer-vetting-framework.md and diagnostic-system.md.
Boundary markers: There are no explicit boundary markers or instructions in the provided files to treat analyzed external content strictly as data or to ignore embedded instructions within that content.
Capability inventory: The skill is authorized to perform file operations, specifically writing to profiles within the ./brand-brain/ directory and logging data to learnings.jsonl as seen in SKILL.md.
Sanitization: No specific sanitization or filtering procedures for ingested natural language data are documented in the skill's reference materials.

afa-influencer