The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a workflow that requires the agent to process data from external MCP resources and tool search results to determine its actions.
Ingestion points: External documentation is fetched via notis_fetch from notis://docs/* URIs, and tool schemas are retrieved through notis_describe_tools.
Boundary markers: The skill relies on structured JSON Schema validation rather than prompt-level text delimiters.
Capability inventory: The agent possesses capabilities to read (notis_execute_read) and write (notis_execute_write) data based on the processed external schemas.
Sanitization: The skill mandates that the agent must validate all calls using notis_validate_calls before execution and specifically forbids inventing arguments or tool names.
[SAFE]: No obfuscation, persistence mechanisms, or unauthorized remote code execution patterns were found. The skill includes proactive security measures, such as instructions to avoid placing secrets in search queries and respecting server-side execution lanes.

notis-mcp