The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions include a directive to read .agents/marketing-context.md if it exists. This represents an ingestion point for external data into the agent's context. However, the skill explicitly includes safety guardrails, such as requiring approval before posting or committing funds, and prohibitions against deceptive practices like impersonation or hiding sponsorships. The risk is minimized by the lack of automated execution capabilities in the provided instructions.

social-community