
from ai-factory921
Set up agent project context: analyze tech stack, install or generate skills, and configure MCP servers with mandatory security scanning for external skills.
AI Factory (/aif) automates project onboarding for agent workflows. It analyzes repository metadata and tech stack, recommends and installs existing skills, generates project-specific skills when needed, and configures MCP servers required for runtime integrations. A key responsibility is enforcing a two-level security scanning process for any external skill: an automated Python scanner followed by a semantic human review to catch prompt-injection or dangerous behaviors.
Run /aif when starting a new project, onboarding an existing repository to agent workflows, or whenever you need a repeatable, secure agent context configured (skills, MCP, AGENTS.md, and project-level rules). Use it at project setup, before running generated or external skills in production, and whenever you add third-party skills.
Designed for agents that can run shell and python helpers and manage MCP servers (Claude Code, Copilot/Codex-style agents, Cursor).
AI Factory (aif) is a comprehensive project onboarding skill that analyzes tech stacks, installs/generates skills, configures MCP servers, and includes mandatory security scanning for external skills. No bundled scripts to test — static analysis only. Very detailed SKILL.md with clear multi-mode workflows, language resolution, and config management via helper scripts.
Well-structured onboarding skill with thoughtful security-first approach (mandatory scanning of external skills). The inline SKILL.md is quite long (~900 lines) which impacts architecture score — moving more content to references/ would improve it. No security concerns found.