The Agent Skills Directory

[SAFE]: The skill is a meta-instruction framework designed to help agents manage project context through modular documentation; no malicious patterns were detected.
[COMMAND_EXECUTION]: The skill guides the agent to create a directory structure (including skills/ and docs/runbooks/) and write markdown files to scaffold the agentic workflow system.
[PROMPT_INJECTION]: The workflow involves processing untrusted codebase data (via git diff) to draft documentation and project plans. This creates a potential surface for indirect prompt injection if a repository contains malicious code comments, although the risk is limited to documentation generation and is managed by explicit review gates.
Ingestion points: Procedure steps in SKILL.md and workflows/document/SKILL.md that read diffScope or project files.
Boundary markers: No specific delimiters or "ignore" instructions are currently included for separating codebase content from the agent's instructions.
Capability inventory: The agent performs filesystem writes and interacts with the user-Linear MCP server to manage tasks.
Sanitization: The procedures do not explicitly define sanitization steps for the data retrieved from the repository before it is used in documentation.

workflow-blueprint