CLI for AI
Robo.js CLI commands produce structured JSON output that AI coding tools can parse and reason over programmatically.
The Robo.js CLI includes commands designed for AI tool consumption. When invoked with --json, these commands produce structured output that AI agents can parse and reason over — no screen-scraping or regex needed.
Skills reference these CLI commands in their workflows. For example, the robo-debug skill uses robo inspect --json and robo logs --json as its primary diagnostic tools.
robo inspect
Display your project's structure, plugins, routes, hooks, configuration, and environment variable health.
npx robo inspect --jsonOptions
| Flag | Description |
|---|---|
--json, -j | Output as structured JSON |
--mode, -m | Which mode manifest to inspect (development, production, or custom) |
Without --json, the command prints a human-readable summary.
JSON output schema
{
"project": {
"name": "my-bot",
"version": "1.0.0",
"language": "typescript",
"roboVersion": "0.11.0",
"mode": "production",
"buildTime": "2025-03-15T10:30:00.000Z"
},
"plugins": [
{
"name": "@robojs/discordjs",
"version": "0.11.0",
"namespace": "discord",
"routes": ["commands", "events", "context", "middleware"],
"hooks": ["start", "stop", "prepare", "hmr"]
}
],
"routes": {
"commands": ["ping", "help", "admin ban"],
"events": ["messageCreate", "ready"],
"api": ["/api/health", "/api/users"],
"context": ["user/Report"],
"middleware": ["logger"]
},
"hooks": ["start", "stop", "prepare"],
"config": {
"type": "robo",
"namespace": "my-bot"
},
"env": {
"summary": { "total": 5, "set": 4, "empty": 0, "missing": 1 },
"missing": ["OPENAI_API_KEY"],
"satisfied": ["DISCORD_TOKEN", "DATABASE_URL", "NODE_ENV", "PORT"]
}
}Key fields
| Field | Description |
|---|---|
project | Project name, version, language, Robo version, build mode, and build time |
plugins | Array of installed plugins with their namespaces, routes, and hooks |
routes | All registered routes grouped by type (commands, events, api, context, middleware) |
hooks | Lifecycle hooks present in the project |
config | Merged project configuration |
env | Environment variable health: total count, how many are set, empty, or missing |
The env.missing array is one of the most useful fields for diagnostics. Missing environment variables are a common source of runtime errors.
robo logs
Query, filter, and stream local log files from .robo/logs/.
npx robo logs --json --level errorOptions
| Flag | Description |
|---|---|
--json, -j | Output as NDJSON (one JSON object per line) |
--level, -l | Filter by minimum log level (trace, debug, info, warn, error) |
--source, -p | Filter by source/plugin name (case-insensitive substring match) |
--grep, -g | Filter log messages by text or regex pattern |
--mode, -m | Which mode logs to read (development, production, or custom) |
--limit, -n | Maximum number of entries to output |
--tail, -t | Watch for new log entries in real time (stream mode) |
NDJSON output format
Each line is a standalone JSON object:
{"timestamp":"10:15:32.000","level":"error","source":"discordjs","message":"Failed to connect to gateway"}
{"timestamp":"10:15:33.100","level":"warn","source":"api","message":"Rate limit approaching for /api/users"}
{"timestamp":"10:15:34.200","level":"info","source":null,"message":"Reconnecting..."}| Field | Type | Description |
|---|---|---|
timestamp | string | Log entry timestamp |
level | string | Log level (trace, debug, info, warn, error, event, ready, wait) |
source | string | null | Plugin or logger fork name (e.g., discordjs, api, ai) |
message | string | The log message content |
Filter examples
# Recent errors only
npx robo logs --json --level error
# Logs from the Discord plugin
npx robo logs --json --source discordjs
# Search for connection issues
npx robo logs --json --grep "ECONNREFUSED"
# Combine filters: warnings from the API layer, last 20 entries
npx robo logs --json --level warn --source api --limit 20
# Stream new errors in real time
npx robo logs --json --level error --tailLog level priority
Levels are ordered by priority. The --level filter shows entries at or above the specified level:
| Level | Priority | Description |
|---|---|---|
trace | 0 | Fine-grained debugging |
debug | 1 | Development diagnostics |
info | 2 | General information |
wait | 3 | Waiting/loading indicators |
other | 4 | Standard output |
event | 5 | Event-driven messages |
ready | 6 | Startup/ready signals |
warn | 7 | Warnings |
error | 8 | Errors |
robo skills
Manage AI coding skills from plugins.
| Command | Description |
|---|---|
npx robo skills | List installed and available skills |
npx robo skills list | List installed and available skills |
npx robo skills install --all | Install from all registered plugins |
npx robo skills install <plugin> | Install from a specific plugin |
npx robo skills remove <plugin> | Remove skills from a plugin |
npx robo skills update | Update all installed skills |
npx robo skills update <plugin> | Update skills from a specific plugin |
Install options
| Flag | Description |
|---|---|
--all, -a | Install from all registered plugins |
--force, -f | Overwrite existing skills from other plugins |
--yes, -y | Skip confirmation prompts |
--silent, -s | Suppress output |
--verbose, -v | Show debug information |
How AI tools use these
The workflow is circular: skills reference CLI commands, CLI commands produce structured output, and AI tools reason over that output.
- A skill (e.g.,
robo-debug) instructs the AI tool to runnpx robo inspect --json - The CLI reads manifest files and outputs structured JSON
- The AI tool parses the JSON to understand project topology
- The tool runs
npx robo logs --json --level errorto check for issues - The tool cross-references errors with the project structure and suggests fixes
This pattern means AI tools can diagnose issues, understand your project, and generate correct code without relying on file-reading heuristics alone.