The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data provided by the user to identify and map stakeholders.
Ingestion points: The skill instructions in SKILL.md explicitly direct the agent to read user-provided files such as organizational charts, project briefs, and team rosters.
Boundary markers: There are no boundary markers or instructions to the agent to ignore any natural language instructions that might be embedded within the source files.
Capability inventory: The skill instructions conclude with a command for the agent to perform a file system operation: "Save the stakeholder map as a markdown document."
Sanitization: The skill does not include any validation, escaping, or filtering logic to ensure that instructions hidden in the input files are not executed by the agent.

stakeholder-map