The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided content such as business plans, strategic decisions, and meeting agendas for the purpose of adversarial analysis. While this is the intended function of the skill, it creates a surface for indirect prompt injection where instructions embedded in the user data could influence the agent. • Ingestion points: User-provided inputs via commands such as /em:challenge, /em:board-prep, and /em:hard-call in SKILL.md. • Boundary markers: No specific delimiters or boundary markers are defined to isolate user-supplied data from the agent instructions. • Capability inventory: The skill utilizes local Python scripts (decision_matrix_scorer.py and stakeholder_mapper.py) for data processing; no network access or file-system write capabilities were detected in the provided code. • Sanitization: The skill relies on platform-level safety filters when processing user-supplied text as it contains no specific sanitization logic.

executive-mentor