ChronoAIProject

agents

Meta-skill for AI agents working with Chrono CLI. Provides navigation to available skills and MCP tools. Use when starting work on Chrono-related tasks to understand available capabilities, or when unsure which skill to use for a specific task.

ChronoAIProject 0 Updated 4mo ago
GitHub

Install

npx skillscat add chronoaiproject/chrono-cli/agents

Install via the SkillsCat registry.

SKILL.md

Agent Instructions

Available Skills

Chrono CLI provides these skills for AI agents:

  • check-setup - Verify project structure before deployment
  • deploy - Deploy current project to the platform
  • restart - Rolling restart backend deployment
  • chrono - Chrono CLI command reference

Skill Selection Guide

User Request Use Skill
"Deploy my project" deploy
"Is my project ready to deploy?" check-setup
"Restart the pods" restart
"How do I use chrono CLI?" chrono
"Setup MCP integration" chrono

Preferred Behavior

  • Reuse existing skills instead of re-implementing logic
  • Read skill files before taking action to understand the proper workflow
  • Ask before creating new skills - the existing skills cover most use cases

Critical: Environment Variables

Before creating a pipeline, you MUST:

  1. Read all .env files in the project:

    • Root: .env, .env.local, .env.production
    • Frontend: frontend/.env*
    • Backend: backend/.env*
  2. Classify variables:

    • frontendEnvVars - public prefixes (NEXT_PUBLIC_*, VITE_*, etc.) - baked into bundle
    • backendEnvVars - non-sensitive backend config (PORT, LOG_LEVEL, etc.)
    • secrets - sensitive data (API_KEY, JWT_SECRET, passwords) - stored securely
  3. Present detected variables to the user in a clear table format

  4. Ask user to confirm or modify before proceeding

  5. Wait for user confirmation - never auto-create pipelines without consent

Note: Backend env vars are injected via Kubernetes secrets at runtime (secure).

MCP Tools

The platform provides MCP tools for direct integration:

User:

  • list_teams - List all teams the user belongs to

GitHub & Projects:

  • check_github_connection - Check GitHub OAuth status
  • start_github_device_flow - Start GitHub device flow authorization
  • poll_github_device_flow - Poll GitHub device flow status
  • list_github_repos - List GitHub repositories
  • list_github_branches - List branches for a GitHub repository
  • detect_repo_type - Detect application type and structure
  • list_projects - List all projects
  • create_project - Create a new project
  • get_project - Get details of a specific project

Pipelines:

  • list_pipelines - List all pipelines (filtered by project/environment/status)
  • create_pipeline - Create CI/CD pipeline (supports frontendEnvVars, backendEnvVars, secrets)
  • get_pipeline - Get details of a specific pipeline including config
  • update_pipeline - Update pipeline config (triggers auto-restart)

Pipeline Runs:

  • trigger_pipeline_run - Trigger a new pipeline run
  • get_run_status - Get current status of a pipeline run
  • list_runs - List recent runs for a pipeline
  • cancel_run - Cancel a running or pending pipeline run

Deployments:

  • restart_deployment - Rolling restart
  • get_deployment_status - Get deployment status
  • get_deployment_logs - Get logs from pods (all pods with labels)
  • get_pod_env_vars - Get env vars from pods (secrets masked)

Prefer using MCP tools over shell commands when available.