The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data fetched from Notion, creating a surface for potential indirect prompt injection attacks.
Ingestion points: The workflow in SKILL.md uses Notion:notion-fetch to retrieve requirements, acceptance criteria, and constraints from Notion specification pages.
Boundary markers: The instructions do not define specific delimiters (such as XML tags or triple backticks) or explicit "ignore instructions" warnings for the content fetched from Notion.
Capability inventory: The skill has the capability to create and update Notion pages (Notion:notion-create-pages, Notion:notion-update-page) and provide shell commands to the user, allowing potentially injected instructions to influence subsequent actions.
Sanitization: No explicit validation or filtering of the fetched Notion content is specified before the data is parsed and used for implementation planning.
[COMMAND_EXECUTION]: The skill provides instructions for managing the environment using the codex CLI.
Step 0 of the workflow guides the agent to inform the user how to run commands such as codex mcp add, codex --enable rmcp_client, and codex mcp login to configure the Model Context Protocol (MCP) environment.
[EXTERNAL_DOWNLOADS]: The skill references an external URL for the Notion MCP server configuration.
It instructs the addition of the Notion MCP from https://mcp.notion.com/mcp, which is an official endpoint belonging to a well-known service.

notion-spec-to-implementation