The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it incorporates external research data and user-supplied context into its persona-based simulation process.
Ingestion points: The agent processes user questions, local project files, and the output from external research tools like deep-research and watch-video (SKILL.md).
Boundary markers: The skill includes a simulation disclaimer at the top of the output and uses a structured markdown format for responses, which provides some structural separation but does not prevent the agent from obeying instructions embedded in the data (SKILL.md).
Capability inventory: The skill is capable of reading and writing files within the ~/.config/makerskills/ directory and executing external research and decision-making tools (SKILL.md).
Sanitization: The instructions do not define any specific methods for sanitizing or escaping the content retrieved from external sources before it is processed by the agent.
[COMMAND_EXECUTION]: The skill requires the agent to execute various tools to gather recent data and perform deep research, including deep-research, watch-video, and /last30days (SKILL.md).

maker-council