What problem does it solve? Agent skills in a repository need a single, committed source of truth for validation commands, tracker integration, label taxonomy, and QA gates. Without it, every skill fails or improvises its own settings. This Skill inspects the repository, asks a few questions, and writes .ai/agentic.config.json plus the supporting docs every other skill reads. ## Core Features & Use Cases - Auto-detected configuration: Resolves the default branch, detects validation commands from package.json scripts, Makefiles, or language conventions (Cargo, Go, Python), and writes the committed .ai/agentic.config.json schema. - Tracker and browser descriptor installation: Installs shipped descriptors for GitHub, Linear, or Jira trackers and agent-browser or Playwright browser providers, with refresh/merge/keep protection for existing files. - Project doc generation: Generates missing SDLC.md, AGENTS.md (with a task-routing table), CODE_REVIEW.md, and BACKWARD_COMPATIBILITY.md, each derived from the actual repository and never overwriting existing files. - Use Case: After installing a collection of agent workflow skills into a fresh repository, run this Skill once so every other skill can load its config, tracker operations, and label taxonomy instead of stopping with a missing-config error. ## Quick Start Ask the agent to run the setup-agent-pipeline skill to configure this repository's agentic pipeline, or pass --defaults to write the auto-detected config without questions.
