Home / Documentation / Commands

MARIA OS Commands

Docs snapshot: MARIA-LITE v11.0.0 · Node: v20.20.2
Generated: 2026-04-17T10:31:51.273Z
Source: shipped src-lite package + synced command reference snapshot
Commands
/account
Show account info (debug diagnostics with MARIA_DEBUG=1).
USAGE
/account
/analytics
Cross-product analytics dashboard — command execution, multi-repo git, server usage, billing, calendar/email.
USAGE
/analytics
EXAMPLES
/analytics --since "2026-02-01"
    /analytics --json
    /analytics --no-google --no-billing
/api-secret
Manage local API secrets (encrypted on disk).
USAGE
/api-secret --service <id> [--list] [--required]
    [/api-secret --service <id> --secret <key=value>]
    [/api-secret --service <id> --secrets <json_object>]
    [/api-secret --service <id> --openapi <https_url> --required]
    [--json]
EXAMPLES
/api-secret --service freee --required
    /api-secret --service freee --openapi https://.../openapi.yml --required
    /api-secret --service freee --secret "oauth_access_token=..."
    /api-secret --service freee --secrets "{"oauth_access_token":"..."}"
/apollo
Apollo GTM wrapper for prospecting, enrichment, and outbound execution APIs.
USAGE
/apollo setup [--openapi <url>] [--docs-url <url>] [--display-name <name>] [--secret <key=value>] [--secrets <json>] [--plan-only] [--json]
  /apollo actions [--action <id>] [--json]
  /apollo run --action <id> [--input <json>] [--input-file <path>] [--json]
  /apollo connect [--status] [--disconnect] [--action <id>] [--json]
  /apollo secret [--list] [--required] [--secret <key=value>] [--secrets <json>] [--json]
  /apollo status [--json]
  Hint: Use `/apollo setup` to register the service catalog, then `/apollo secret --required` for API keys.
  Auth: Apollo integrations are commonly API-key based; use `/apollo secret --secret "<key>=<value>"`.
/assert
Deterministic assertions for flows (existence/validity checks).
USAGE
/assert existsAll=path1,path2 [existsAny=...] [missingAll=...] [missingAny=...] [jsonValidAll=...] [jsonValidAny=...] [base=root|cwd]
/auto-dev
Auto-dev (beta): generate no-email-gated extension commands under maria-ext-beta/ (does NOT edit MARIA core code).
USAGE
/auto-dev gen --goal "..." [--command-id <id>] [--sample] [--plan-only] [--json]
  
  Notes:
  - Default output is local desktop: ~/.maria/desktop/artifacts/maria-ext-beta/packs/...
  - Use --sample to write into repo sample packs: maria-ext-beta/packs/...
  - It does not modify MARIA core code. Generated commands run via a generic beta extension runtime.
/billing
Open billing portal / manage plan flow.
USAGE
/billing [--no-open] [--json]
/billing-pl
Billing P&L analysis with subcommands: timesheet, alert, approve, kpi. Default: P&L query.
USAGE
/billing-pl [subcommand] [args] [--flags]
    (no subcmd)  [projectId] [month] [--cumulative] [--mrr]
    timesheet    [month]           — Timesheet summary
    alert        [--all]           — Agent alerts
    approve      <decision-id>     — Approve a proposal
    kpi          [month]           — KPI dashboard
/blog
Generate and sync blogs. Applies by default.
USAGE
/blog generate [--date YYYYMMDD] [--since <YYYYMMDD|ISO>] [--slot 1|2|3] [--category benchmark|decision|reflection] [--force] [--replace]
  /blog approve-publish [--revoke] (writes ~/.maria/desktop/artifacts/blog/approval.flags.json)
  /blog sync [--date YYYYMMDD] [--slot 1|2|3] [--project <id>] [--publish] [--in <dir>] [--reauth-adc]
/bulk-register
Bulk-register API catalogs from multiple OpenAPI sources (applies by default; use --plan-only to preview).
USAGE
/bulk-register [--source all|api-guru|openapisearch] [--recompile] [--offset 0] [--match <substring>] [--format yaml|json]
    [--limit <n>] [--concurrency 4] [--update] [--plan-only] [--json]
    (shorthand flags: --api-guru, --openapisearch)
/caio
Human-AI architecture and UX design support as CAIO (MARIA).
USAGE
/caio [--profile <id>] [--provider <lmstudio|ollama|vllm>] [--model <name>] [--inputs <json|@file>] [--use-latest off|on|auto] [--auto-rerun off|on|auto] [--json] "Your request"
/caio-clone
CAIO Clone (lite): decision-structure replay engine for AI management boundaries with audit and local-only contract. Writes artifacts under artifacts/caio-clone by default.
USAGE
/caio-clone init
  /caio-clone boundary template [--out <file>] [--replace]
  /caio-clone boundary set --file <boundary.json>
  /caio-clone profile template [--out <file>] [--replace]
  /caio-clone profile set --file <profile.json>
  /caio-clone profile show
  /caio-clone policy d3 template [--out <file>] [--replace]
  /caio-clone policy d3 set --file <d3-policy.json>
  /caio-clone policy review-slo template [--out <file>] [--replace]
  /caio-clone policy review-slo set --file <review-slo.json>
  /caio-clone review init
  /caio-clone review enqueue --file <candidate.json> [--reviewer <name>]
  /caio-clone review approve|discard|edit --id <candidateId> --reviewer <name> [--notes <text>]
  /caio-clone ingest template [--out <file>] [--replace]
  /caio-clone ingest --file <decisionlog.json>
  /caio-clone query template [--out <file>] [--replace]
  /caio-clone query --file <decisionrequest.json>
  
  Typical flow:
    1) /caio-clone init
    2) /caio-clone boundary template --replace (or boundary set --file ...)
    3) /caio-clone ingest template --replace (or ingest --file ...)
    4) /caio-clone query template --replace (or query --file ...)
    5) /caio-clone query --file <decisionrequest.json>
  
  Notes:
    - Query writes artifacts under artifacts/caio-clone/runs/<runId>/ by default.
/calculator
Deterministic math calculator: arithmetic, combinatorics, statistics, calculus, and more.
USAGE
/calculator <expression> [--mode basic|advanced|stats] [--data "1,2,3"] [--derivative-at <x>] [--integral-from <a> --integral-to <b>]
/ceo
Executive decision support (capital policy, portfolio, org design) as CEO (MARIA).
USAGE
/ceo [--profile <id>] [--provider <lmstudio|ollama|vllm>] [--model <name>] [--inputs <json|@file>] [--use-latest off|on|auto] [--auto-rerun off|on|auto] [--json] "Your question"
/ceo-clone
CEO Clone (commercial v1.2): decision-structure replay engine with local-only contract, audit, and bounded output. Writes artifacts under artifacts/ceo-clone by default.
USAGE
/ceo-clone init
  /ceo-clone boundary template [--out <file>] [--replace]
  /ceo-clone boundary set --file <boundary.json>
  /ceo-clone profile template [--out <file>] [--replace]
  /ceo-clone profile set --file <profile.json>
  /ceo-clone profile show
  /ceo-clone policy d3 template [--out <file>] [--replace]
  /ceo-clone policy d3 set --file <d3-policy.json>
  /ceo-clone policy review-slo template [--out <file>] [--replace]
  /ceo-clone policy review-slo set --file <review-slo.json>
  /ceo-clone review init
  /ceo-clone review enqueue --file <candidate.json> [--reviewer <name>]
  /ceo-clone review approve|discard|edit --id <candidateId> --reviewer <name> [--notes <text>]
  /ceo-clone ingest template [--out <file>] [--replace]
  /ceo-clone ingest --file <decisionlog.json>
  /ceo-clone query template [--out <file>] [--replace]
  /ceo-clone query --file <decisionrequest.json>
  
  Typical flow (minimum):
    1) /ceo-clone init
    2) /ceo-clone boundary template --replace (or boundary set --file ...)
    3) Ingest decision logs (choose one):
       - /ceo-clone ingest template --replace; then /ceo-clone ingest --file <decisionlog.json>
       - /ceo-clone review init; /ceo-clone review enqueue --file <candidate.json>; /ceo-clone review approve --id <candidateId> --reviewer <name>
    4) /ceo-clone query template --replace; then /ceo-clone query --file <decisionrequest.json>
  
  Optional setup (recommended):
    - /ceo-clone profile set --file <profile.json>
    - /ceo-clone policy d3 set --file <d3-policy.json>
    - /ceo-clone policy review-slo set --file <review-slo.json>
  
  Notes:
    - Query writes artifacts under artifacts/ceo-clone/runs/<runId>/ by default.
/chat
Chat via api.maria-code.ai (ai-proxy compatible).
USAGE
/chat "message"
/clay
Clay GTM wrapper for enrichment, research tables, and workflow automation APIs.
USAGE
/clay setup [--openapi <url>] [--docs-url <url>] [--display-name <name>] [--secret <key=value>] [--secrets <json>] [--plan-only] [--json]
  /clay actions [--action <id>] [--json]
  /clay run --action <id> [--input <json>] [--input-file <path>] [--json]
  /clay connect [--status] [--disconnect] [--action <id>] [--json]
  /clay secret [--list] [--required] [--secret <key=value>] [--secrets <json>] [--json]
  /clay status [--json]
  Hint: Run `/clay setup`; if discovery misses docs, re-run with `--docs-url` or `--openapi`.
  Auth: Clay is usually API-key based; configure credentials with `/clay secret`.
/code
Generate code changes. Applies by default; use --plan-only to preview without writes.
USAGE
/code "request" [--plan-only] [--files "path1,path2"]
    [--in-place] [--new-workspace] [--workspace "relative/path|/abs/path"] [--set-repos-dir "/abs/path"]
    [--allow-new-top-level "dir1,dir2"] [--allow-new-dirs "a/b,c/d"]
/code-quality
AI-agent-oriented code quality audit for a repository: searchability, extension inference, docs tightness, and architectural discipline.
EXAMPLES
/code-quality "."
    /code-quality --path "./apps/admin" --repo "owner/repo" --gh-issue
    /code-quality --repo "owner/repo" --docx --pdf --json
/competitors
Competitive analysis: research domestic/global competitors, generate executive report (DOCX+PDF).
EXAMPLES
/competitors "AI-powered project management tool" --domestic --global
    /competitors --in @product-desc.md --domestic
/connect
Connect OAuth for a registered API service (stores refreshable tokens locally).
USAGE
/connect --service <id> [--action <actionId>] [--status] [--disconnect] [--secret-key <key>] [--authorization-url <url>] [--token-url <url>] [--pkce on|off] [--port 53682] [--json]
  Notes:
  - Requires OAuth app config via /api-secret for the service:
    /api-secret --service <id> --secret "oauth_client_id=..."
    /api-secret --service <id> --secret "oauth_client_secret=..."  (optional)
    /api-secret --service <id> --secret "<secretKey>_authorization_url=https://..."
    /api-secret --service <id> --secret "<secretKey>_authorize_url=https://..."  (alias)
    /api-secret --service <id> --secret "<secretKey>_token_url=https://..."
/constitution
Manage MARIA's organizational constitution, current phase, and drift detection against recent behavior.
USAGE
/constitution
  /constitution init [--name <org>] [--force]
  /constitution review [--json]
  /constitution phase [--set survival|accumulation|liberation] [--json]
  /constitution drift [--limit 40] [--json]
  /constitution set-purpose --id <id_optional> --principle "..." [--evaluation "..."]
  /constitution set-origin --value "..."
  /constitution add-principle --title "..." --body "..." [--priority 100] [--tags "trust,quality"]
  /constitution add-forbidden --pattern "..." --reason "..." [--scope command|text|both] [--severity warn|block]
  /constitution set-stakeholder --name users --weight 100 [--description "..."]
  /constitution add-tension --when "..." --prefer "..." --because "..."
/coo
Operations design, scaling, progress management, and incident response as COO (MARIA).
USAGE
/coo [--profile <id>] [--provider <lmstudio|ollama|vllm>] [--model <name>] [--inputs <json|@file>] [--use-latest off|on|auto] [--auto-rerun off|on|auto] [--json] "Your request"
/coo-clone
COO Clone (lite): decision-structure replay engine for operations boundaries with audit and local-only contract. Writes artifacts under artifacts/coo-clone by default.
USAGE
/coo-clone init
  /coo-clone boundary template [--out <file>] [--replace]
  /coo-clone boundary set --file <boundary.json>
  /coo-clone profile template [--out <file>] [--replace]
  /coo-clone profile set --file <profile.json>
  /coo-clone profile show
  /coo-clone policy d3 template [--out <file>] [--replace]
  /coo-clone policy d3 set --file <d3-policy.json>
  /coo-clone policy review-slo template [--out <file>] [--replace]
  /coo-clone policy review-slo set --file <review-slo.json>
  /coo-clone review init
  /coo-clone review enqueue --file <candidate.json> [--reviewer <name>]
  /coo-clone review approve|discard|edit --id <candidateId> --reviewer <name> [--notes <text>]
  /coo-clone ingest template [--out <file>] [--replace]
  /coo-clone ingest --file <decisionlog.json>
  /coo-clone query template [--out <file>] [--replace]
  /coo-clone query --file <decisionrequest.json>
  
  Typical flow:
    1) /coo-clone init
    2) /coo-clone boundary template --replace (or boundary set --file ...)
    3) /coo-clone ingest template --replace (or ingest --file ...)
    4) /coo-clone query template --replace (or query --file ...)
    5) /coo-clone query --file <decisionrequest.json>
  
  Notes:
    - Query writes artifacts under artifacts/coo-clone/runs/<runId>/ by default.
/cpo
Product vision, prioritization, UX, and roadmap design as CPO (MARIA).
USAGE
/cpo [--profile <id>] [--provider <lmstudio|ollama|vllm>] [--model <name>] [--inputs <json|@file>] [--use-latest off|on|auto] [--auto-rerun off|on|auto] [--json] "Your request"
/cpo-clone
CPO Clone (lite): decision-structure replay engine for product boundaries with audit and local-only contract. Writes artifacts under artifacts/cpo-clone by default.
USAGE
/cpo-clone init
  /cpo-clone boundary template [--out <file>] [--replace]
  /cpo-clone boundary set --file <boundary.json>
  /cpo-clone profile template [--out <file>] [--replace]
  /cpo-clone profile set --file <profile.json>
  /cpo-clone profile show
  /cpo-clone policy d3 template [--out <file>] [--replace]
  /cpo-clone policy d3 set --file <d3-policy.json>
  /cpo-clone policy review-slo template [--out <file>] [--replace]
  /cpo-clone policy review-slo set --file <review-slo.json>
  /cpo-clone review init
  /cpo-clone review enqueue --file <candidate.json> [--reviewer <name>]
  /cpo-clone review approve|discard|edit --id <candidateId> --reviewer <name> [--notes <text>]
  /cpo-clone ingest template [--out <file>] [--replace]
  /cpo-clone ingest --file <decisionlog.json>
  /cpo-clone query template [--out <file>] [--replace]
  /cpo-clone query --file <decisionrequest.json>
  
  Typical flow:
    1) /cpo-clone init
    2) /cpo-clone boundary template --replace (or boundary set --file ...)
    3) /cpo-clone ingest template --replace (or ingest --file ...)
    4) /cpo-clone query template --replace (or query --file ...)
    5) /cpo-clone query --file <decisionrequest.json>
  
  Notes:
    - Query writes artifacts under artifacts/cpo-clone/runs/<runId>/ by default.
/cto
CTO Engineering OS guidance: technical debt, SLO/SLA, quality, throughput.
USAGE
/cto [--profile <id>] [--use-latest off|on|auto] [--auto-rerun off|on|auto] [--json] "Your question"
/cto-clone
CTO Clone (commercial v1.2): decision-structure replay engine for engineering/tech boundaries with audit and local-only contract. Writes artifacts under artifacts/cto-clone by default.
USAGE
/cto-clone init
  /cto-clone boundary template [--out <file>] [--replace]
  /cto-clone boundary set --file <boundary.json>
  /cto-clone profile template [--out <file>] [--replace]
  /cto-clone profile set --file <profile.json>
  /cto-clone profile show
  /cto-clone policy d3 template [--out <file>] [--replace]
  /cto-clone policy d3 set --file <d3-policy.json>
  /cto-clone policy review-slo template [--out <file>] [--replace]
  /cto-clone policy review-slo set --file <review-slo.json>
  /cto-clone review init
  /cto-clone review enqueue --file <candidate.json> [--reviewer <name>]
  /cto-clone review approve|discard|edit --id <candidateId> --reviewer <name> [--notes <text>]
  /cto-clone ingest template [--out <file>] [--replace]
  /cto-clone ingest --file <decisionlog.json>
  /cto-clone query template [--out <file>] [--replace]
  /cto-clone query --file <decisionrequest.json>
  
  Typical flow:
    1) /cto-clone init
    2) /cto-clone boundary template --replace (or boundary set --file ...)
    3) /cto-clone ingest template --replace (or ingest --file ...)
    4) /cto-clone query template --replace (or query --file ...)
    5) /cto-clone query --file <decisionrequest.json>
  
  Notes:
    - Query writes artifacts under artifacts/cto-clone/runs/<runId>/ by default.
/cxo --members "ceo,coo,caio" "Decide go/no-go with options." --json
Get a structured decision view across multiple executive roles.
/daily-insights
Synthesize daily data, news, and papers into actionable insights with executable MARIA commands.
USAGE
/daily-insights "focus area" [--topics "AI, ML" | --topics 3] [--out <path>] [--json] [--discord [channel]]
/daily-news
Search and curate relevant news based on your professional activity. Optionally post to Discord.
USAGE
/daily-news "query" [--topics "AI, ML" | --topics 3] [--days <n>] [--articles <n>] [--out <path>] [--json] [--discord [channel]]
/daily-papers
Search and curate relevant academic papers based on your professional activity. Optionally post to Discord.
USAGE
/daily-papers "query" [--topics "AI, ML" | --topics 3] [--papers <n>] [--out <path>] [--json] [--discord [channel]]
/daily-report
Generate a daily activity report from GitHub, Google Calendar, and Gmail. Optionally post to Discord.
USAGE
/daily-report [--since <YYYY-MM-DD>] [--days <n>] [--format <markdown|slack|plain>] [--max-repos <n>] [--out <path>] [--json] [--discord [channel]]
/deps
Installer helpers for required tools. Applies by default; use --plan-only to preview.
USAGE
NPM global installers:
    /deps npm-global --pkg "md-to-pdf" [--bin md-to-pdf] [--plan-only] [--json]
    /deps npm-global --pkg "@scope/pkg" [--version "1.2.3"] [--bin mycmd] [--plan-only] [--json]
  
  Notes:
  - Default is apply. Use --plan-only to preview without installing.
  - npm install always uses --legacy-peer-deps.
/desire
Desire (lite): set and manage the system's active desires (knowledge/creation) used by /self-recognition and /evolve.
USAGE
Notes:
  - default is local per-user store under artifacts/desires/desires.json
  - add --sample to use the repo sample store under src-lite/desires/desires.json
  
    /desire list [--sample] [--json]
    /desire set "<topic>" [--sample] [--raw] [--kind knowledge|creation] [--json]
    /desire remove <id> [--sample] [--json]
    /desire clear [--sample] [--json]
/desktop
Start local Desktop OS UI in your browser (app launcher + window manager + file manager + parallel execution).
USAGE
/desktop [--host 127.0.0.1] [--port 57597] [--no-open] [--json]
  
  Notes:
  - Default binds to 127.0.0.1 (local-only).
  - If --port is omitted, it tries 57597 and falls back to a random free port if needed.
  - If --port is 0, a random free port is chosen.
  - Use --no-open to avoid opening a browser tab.
  - Unlike /gui, /desktop supports parallel command execution.
  - Desktop files are stored at ~/.maria/desktop/.
/dev-adviser
Code-first developer assessment — Quantity, Architect Quality, and Semantic Quality evaluation with trend analysis.
USAGE
Usage: /dev-adviser <username> [options]
EXAMPLES
/dev-adviser "octocat"
    /dev-adviser "octocat" --days 7
    /dev-adviser "johndoe" --org mycompany --docx --pdf
    /dev-adviser "janedoe" --repo owner/repo --discord
/dev-decision
Analyze AI tool conversation logs (Cursor, Codex, Claude Code) to identify development decision trends and propose next strategy.
USAGE
Usage: /dev-decision [options]
EXAMPLES
/dev-decision
    /dev-decision --weeks 8 --tools "claude,codex"
    /dev-decision --repo "myorg/myrepo" --gh-issue
    /dev-decision --discord --docx --pdf
/develop
Development kickoff (lite): design → tasks → initial handoff bundle. Emits artifacts for humans (/auto-dev job spec draft).
USAGE
/develop "<goal>" [--mode spec|plan|full] [--spec <path>] [--context <path>] [--product <id>] [--json] [--plan-only]
/discord-setup
Configure Discord webhook integration for daily intelligence posting.
USAGE
/discord-setup add --name <channel> --webhook <url> [--username <bot-name>] [--avatar <url>]
  /discord-setup remove --name <channel>
  /discord-setup list
  /discord-setup test [--name <channel>]
  /discord-setup set-default --name <channel>
  /discord-setup set-prompt --key <key> --value <prompt-text>
/docs
Generate proposal meeting docs (talk track, Q&A, prep materials) and export to DOCX+PDF.
EXAMPLES
/docs --in @proposal/source.md --docx --pdf
    /docs --title "ACME proposal notes" "Create talk track for the ACME meeting"
/docs-update
Refresh the managed root README block (applies by default; use --plan-only to preview).
USAGE
Write the managed README block into the repo root README (default):
    /docs-update
  
  Preview without writes:
    /docs-update --plan-only
/doctor
Repo health checks (lite): runs SSOT quality gates (lint:truth + tsc --noEmit + lite boundary check). Add --audit to append repo-wide code-quality + UX audit summary. Optionally asks ai-proxy for a short explanation and next steps.
USAGE
/doctor [--json] [--audit] [--deep-dive] [--lint] [--typecheck] [--boundary] [--test]
  Defaults: lint+typecheck+boundary. Use --audit to append repo-wide code-quality + UX findings, and --deep-dive to add ai-proxy summary (no code changes).
/estimate
Effort estimation with 3-tier coefficients and multi-person scheduling.
EXAMPLES
/estimate "Add OAuth2 SSO" --repo "owner/repo"
    /estimate --in @brief.md --devs 3 --xlsx --pdf
    /estimate "Migrate to PostgreSQL" --repo "org/app" --devs 5 --skills "Alice:backend,Bob:frontend" --xlsx --pdf
/events
Inspect and publish MARIA event bus records, and process escalation TTL notifications.
USAGE
/events list [--limit 20] [--kind <kind>] [--severity info|warn|error|critical] [--topic <topic>]
  /events publish --kind ops.custom --topic my-topic --severity warn --message "..." [--ttl 30m]
  /events escalate [--limit 20]
/evolve
Evolve (lite): apply improvements for KP/Universe based on /self-recognition, or generate a generic evolve plan.
USAGE
1) Apply improvements (recommended):
     /evolve --kp [--direction "..."] [--dev] [--kp-parallel 3] [--max-rounds 3] [--top-k 3] [--max-packs 3] [--json]
     /evolve --kp --synthesis [--dev] [--ndc 3] [--pick 2] [--ndc3 "<csv_optional>"] [--json]
     /evolve --kp --update [--dev] [--max-packs 3] [--max-queries 3] [--top-k 5] [--ndc3 320] [--json]
     /evolve --universe [--sample] [--json]
     /evolve --desire [--json]
     /evolve --skills [--json]
     /evolve --persona [--sample] [--direction "..."] [--max-topics 5] [--json]
     /evolve --kp --universe [--direction "..."] [--dev] [--sample] [--json]
     - Uses /self-recognition gap topics (default 5) and fixes 1 topic per selected domain.
  
  2) Code evolve (beta):
     /evolve --code [--json]
     /evolve --code --issues [--json]
     - Generates a no-email-gated extension command under maria-ext-beta/ via /auto-dev (beta).
     - With --issues: create GitHub issue(s) only (no code generation).
  
  3) Legacy plan mode:
     /evolve "<goal>" [--json]
  4) Status:
     /evolve status [--run <runId>] [--json]
/exam
Answer exam questions: reads exam content, researches as needed, provides structured answers with LaTeX math. Export PDF+DOCX.
EXAMPLES
/exam --in @exams/midterm.pdf "Answer all questions"
    /exam --in @exam.txt --lang ja
    /exam "What is the eigenvalue decomposition of a symmetric matrix?"
/film
Generate a short film from a prompt/material: shared characters+setting, storyboard, per-scene video clips, then concat.
EXAMPLES
/film "A romantic comedy misunderstanding on a space station" --duration 30 --aspect 16:9
    /film --in @docs/brief.md --duration 45 --aspect 9:16 --format mp4
/gcal
Manage Google Calendar events.
USAGE
/gcal list [--max <count>] [--days <n>] [--calendar <calendarId>]
  /gcal get <eventId>
  /gcal create --title "Meeting" --start "2025-03-01T10:00:00" --end "2025-03-01T11:00:00" [--all-day] [--description <text>] [--attendees <emails>]
  /gcal create --title "Day Off" --start "2025-03-01" --end "2025-03-02" --all-day  (all-day: use YYYY-MM-DD format)
  /gcal update <eventId> [--title <text>] [--start <iso>] [--end <iso>] [--all-day] [--description <text>] [--attendees <emails>] [--location <text>]
  /gcal delete <eventId>
  /gcal calendars
  /gcal freebusy --emails <email1,email2> --from <YYYY-MM-DD> --to <YYYY-MM-DD>
  /gcal find-slots --required <email1,email2> [--optional <email3>] [--duration <minutes>] --from <YYYY-MM-DD> --to <YYYY-MM-DD> [--work-start <9>] [--work-end <18>] [--resource <calendarId>] [--capacity <n>] [--building <id>] [--offset <n>] [--buffer-min <minutes>]
  /gcal check --resource <calendarId> --datetime <ISO8601> [--duration <minutes>]
  /gcal check --emails <email1,email2> --datetime <ISO8601> [--duration <minutes>]
  /gcal resources list [--building <buildingId>]
  /gcal resources get <resourceId>
  /gcal resources create --name "Room A" --building <buildingId> --capacity <n> [--type CONFERENCE_ROOM] [--floor <floor>] [--description <text>]
  /gcal resources bulk-import --file <path.csv|json> [--plan-only]
  /gcal buildings list
  /gcal buildings create --name "Main Office" [--floors <n>] [--description <text>]
  /gcal buildings bulk-import --file <path.csv|json> [--plan-only]
/gdrive
Browse and manage Google Drive files.
USAGE
/gdrive list [--max <count>] [--folder <folderId>]
  /gdrive search "<query>" [--max <count>]
  /gdrive info <fileId>
  /gdrive download <fileId> [--out <path>]
  /gdrive read <fileId>         (read Google Docs/Sheets/Slides content)
  /gdrive write <fileId> <text> (append text to a Google Doc)
  /gdrive upload <filePath> [--folder <folderId>] [--name <name>]
  /gdrive tree [--folder <folderId>] [--depth <N>]
/gen-skills
Generate Claude Code-compatible skills (SKILL.md) for MARIA-LITE.
USAGE
Generate one skill:
    /gen-skills --cc --name <skill-name> [--description "..."] [--out <dir>] [--json]
  
  Generate a workflow skill for /self-recognition -> /evolve:
    /gen-skills --cc --name self-recognition-evolve
  
  Install samples into a target dir:
    /gen-skills --cc --sample --out .claude/skills
/gh
Run the system `gh` command (GitHub CLI passthrough).
USAGE
/gh <args...>
  Example: /gh pr status
/git
Run the system `git` command (passthrough).
USAGE
/git <args...>
  Example: /git status -sb
/github-webhook
Configure the inbound GitHub webhook endpoint and optional automatic goal-loop execution.
USAGE
/github-webhook
  /github-webhook setup [--secret "..."] [--events "push,pull_request,issues,issue_comment,workflow_run"] [--goal-loop-auto-fix true|false] [--goal-loop-apply true|false] [--goal-loop-max-topics 5]
  /github-webhook status [--json]
  /github-webhook rotate-secret
  /github-webhook disable
/gmail
Read, send, reply, and event-sync Gmail inbox activity.
USAGE
/gmail inbox [--max <count>]
  /gmail read <messageId>
  /gmail send --to <email> --subject <text> --body <text>
  /gmail reply <messageId> --body <text>
  /gmail search "<query>" [--max <count>]
  /gmail watch sync [--query <gmail_query>] [--max <count>]
  /gmail watch status [--query <gmail_query>]
/gmeet
Create and manage Google Meet meetings.
USAGE
/gmeet create [--title "Meeting title"]
  /gmeet get <spaceId>
  /gmeet list [--max <count>]
  /gmeet recordings <conferenceId> [--max <count>]
  /gmeet transcripts <conferenceId> [--max <count>]
/goal-loop
Event-driven autonomous goal loop: infer domains from trigger context and run self-improve with saved reports.
USAGE
/goal-loop run [--source manual] [--reason "..."] [--event-kind github.push] [--event-topic owner/repo] [--summary "..."] [--auto-fix] [--kp] [--universe] [--code] [--desire] [--skills] [--persona] [--max-topics 5] [--json]
  /goal-loop status [--json]
/gong
Gong GTM wrapper for call review, conversation analytics, and coaching workflows.
USAGE
/gong setup [--openapi <url>] [--docs-url <url>] [--display-name <name>] [--secret <key=value>] [--secrets <json>] [--plan-only] [--json]
  /gong actions [--action <id>] [--json]
  /gong run --action <id> [--input <json>] [--input-file <path>] [--json]
  /gong connect [--status] [--disconnect] [--action <id>] [--json]
  /gong secret [--list] [--required] [--secret <key=value>] [--secrets <json>] [--json]
  /gong status [--json]
  Hint: Use `/gong setup` before `/gong run` so conversation and call endpoints are discoverable.
  Auth: Gong integrations often require API tokens or OAuth depending on the account configuration.
/google-connect
Connect Google Workspace (Gmail, Calendar, Drive, Meet) via OAuth.
USAGE
/google-connect connect [--services gmail,calendar,drive,meet]
  /google-connect status
  /google-connect disconnect
  /google-connect setup --client-id <ID> --client-secret <SECRET>
/gtm
Generate GTM-engineering playbooks and run a local-first revenue operating loop when external GTM SaaS is not connected.
USAGE
Blueprint examples:
    /gtm brief "AI coding platform for enterprise engineering teams"
    /gtm icp --in @docs/product-brief.md --market "Japan + US B2B SaaS"
    /gtm sequence "Vertical AI copilot for support teams" --channels "email,linkedin,partner"
  
  Operational examples:
    /gtm ingest --lead "{"email":"cto@example.com","company":"Example","signals":["funding","ai"]}"
    /gtm ingest --source gmail --query "label:inbox newer_than:14d" --max 25
    /gtm leads
    /gtm pipeline
    /gtm toolkit
    /gtm autopilot --dispatch email
    /gtm proposal --lead <leadId> --amount-usd 5000
    /gtm contract draft --lead <leadId> --amount-usd 5000
    /gtm contract send --lead <leadId> --dispatch email
    /gtm invoice --lead <leadId> --amount-usd 5000
    /gtm collect queue
    /gtm collect remind --lead <leadId> --dispatch email
    /gtm tickets ingest --query "label:inbox newer_than:7d" --max 25
    /gtm tickets reply --ticket <ticketId> --body "We are on it."
    /gtm renew draft --lead <leadId> --amount-usd 5000
    /gtm stage --lead <leadId> --status replied
  
  Modes:
    brief, icp, signals, sequence, score, crm, calls
/gui
Start local GUI in your browser (runs + logs + REPL history + universe inflate graph).
USAGE
/gui [--host 127.0.0.1] [--port 57596] [--no-open] [--json]
  
  Notes:
  - Default binds to 127.0.0.1 (local-only).
  - If --port is omitted, it tries 57596 and falls back to a random free port if needed.
  - If --port is 0, a random free port is chosen.
  - Use --no-open to avoid opening a browser tab.
/help
Show MARIA-LITE commands, usage, and starter guides.
USAGE
/help [command|start|developer|manager|creator|cxo]
/hubspot
HubSpot GTM wrapper for CRM, lifecycle, and automation API flows.
USAGE
/hubspot setup [--openapi <url>] [--docs-url <url>] [--display-name <name>] [--secret <key=value>] [--secrets <json>] [--plan-only] [--json]
  /hubspot actions [--action <id>] [--json]
  /hubspot run --action <id> [--input <json>] [--input-file <path>] [--json]
  /hubspot connect [--status] [--disconnect] [--action <id>] [--json]
  /hubspot secret [--list] [--required] [--secret <key=value>] [--secrets <json>] [--json]
  /hubspot status [--json]
  Hint: Run `/hubspot setup` first, then `/hubspot actions` to inspect available CRM endpoints.
  Auth: HubSpot typically uses OAuth for app installs or private app tokens via `/hubspot secret`.
/image
Generate images via api.maria-code.ai (compatible with /api/v1/image). Saves to outputs/images by default.
USAGE
/image "prompt" [--size 1024x1024] [--format webp|png|jpg] [--count 1..8] [--model <id>] [--seed N] [--ref @a.png,@b.jpg] [--out <dir>] [--path <file>] [--plan-only]
/import-origin
Import original files/folders as-is, build a local origin index for RAG retrieval.
USAGE
/import-origin --path <file_or_dir> [--name <label>] [--w] [--sample] [--max-files 10000] [--json]
  Notes:
  - Default destination: artifacts/origin/<label>/ (and artifacts/origin/index.jsonl)
  - With --sample: src-lite/origin/<label>/ (and src-lite/origin/index.jsonl; included in npm publish)
/investment
Investment simulator with theory-driven allocation, news-sensitive scoring, service selection, and portfolio P/L output.
USAGE
/investment latest [--json]
  /investment profiles [--json]
  /investment history [--profile <name>] [--limit <n>] [--json]
  /investment simulate --capital <amount> --country <country> [--currency <code>] [--goals "short-term-max,FIRE"] [--risk <conservative|balanced|aggressive>] [--months <n>] [--region-focus "Japan,US"] [--industry-focus "AI,Energy"] [--benchmarks "sp500,all-country,nvidia"] [--news-sources "google-news-rss,bing-news-rss"] [--compare-news-sources] [--review-every <hours>] [--trade-every <hours>] [--take-profit] [--take-profit-interval monthly|quarterly|semiannual|annual] [--take-profit-payout <0-100>] [--save-profile <name>] [--discord [channel]] [--mode simulation|paper|live] [--simulation-mode current|past] [--past-window 1m|3m|6m|1y|custom] [--past-start YYYY-MM-DD] [--out <path>] [--json]
/kp-generate
Generate new knowledge packs from web research (default) or local documents (--local), then write YAML packs and rebuild index by default.
USAGE
/kp-generate "<topic>" [--dev] [--max-rounds 10] [--top-k 10] [--max-packs 10] [--json]
  /kp-generate "<topic>" --local --path "<file_or_dir>" [--dev] [--max-packs 10] [--json]
/kp-sync
Sync knowledge packs from `config/knowledge-packs/` into `src-lite/knowledge-packs/` and rebuild the index.
USAGE
/kp-sync <packId|filename|relative/path...> [--plan-only] [--no-rebuild] [--json]
/kp-synthesis
Synthesize new memes/hypotheses by combining multiple NDC domains and selected knowledge packs. Applies by default and rebuilds the index.
USAGE
/kp-synthesis [--dev] [--ndc 3] [--pick 2] [--ndc3 "<csv_optional>"] [--json]
  /kp-synthesis --ndc 3 --pick 1
/kp-update
Scan existing knowledge packs, detect outdated parts, and apply minimal (diff-style) updates. Applies by default and rebuilds the index.
USAGE
/kp-update [--dev] [--max-packs 3] [--max-queries 3] [--top-k 5] [--json]
  /kp-update "<hint_topic>" [--dev] [--max-packs 3] [--json]
/list-api
List available registered APIs (remote if supported; fallback to local cache).
USAGE
/list-api [--limit 50] [--service <id>] [--action <id>] [--search <query>] [--json]
/login
Sign in to MARIA (auth.maria-code.ai).
USAGE
/login [--device] [--force]
/logout
Sign out from MARIA.
USAGE
/logout [--force] [--all]
/manga
Generate manga: 4-panel (one image) or multi-page (default 16 pages). Saves all intermediate plans/prompts/images under ./manga/.
USAGE
/manga "<theme>" [--type 4koma|pages] [--pages N] [--mode oneshot|serial]
    [--in @path1,@path2] [--reuse @path] [--color monochrome|full-color] [--size <WxH>] [--format webp|png|jpg] [--model <id>] [--seed N] [--out manga]
    [--persona <id>]
    --plan-only
/music
Compose music via LLM (ABC + MIDI/WAV/MP3) and save under ./music/.
USAGE
/music "prompt" [--type bgm|song] [--bpm 120] [--time 4/4] [--key <key>] [--bars 64] [--title "..."]
    [--instrument-palette acoustic|raw] [--soundfont /abs/path.sf2] [--persona <id>] [--out music] [--plan-only] [--json]
  
  Notes:
  - Runs immediately by default. Use --plan-only if you want a preview first.
  - WAV rendering prefers synth-js (GM instruments). Fallback is a simple waveform synth.
  - MP3 export uses ffmpeg. If missing, MARIA-LITE will try to auto-install ffmpeg (best-effort) then retry.
  - Instrument palette default is acoustic (to reduce electronic bias). Use --instrument-palette raw to keep model-chosen programs.
  - If no SoundFont is configured, /music will auto-download a default SoundFont on first run.
    - URL: https://archive.org/download/free-soundfonts-sf2-2019-04/airfont_380_final.sf2
    - Cache: ~/.maria/assets/music/airfont_380_final.sf2 (or MARIA_CONFIG_DIR)
/notification-router
Configure notification routing for MARIA event bus events (Discord/Slack/email/log) and test routes.
USAGE
/notification-router
  /notification-router add-discord --name ops-critical --min-severity error [--kinds "lite.command_error,*.failed"] [--topics "self-improve,doctor"] [--channel default]
  /notification-router add-slack --name ops-slack --min-severity warn [--kinds "goal-loop.*,gmail.*"] [--target incoming_webhook_url]
  /notification-router add-email --name founder-mail --min-severity error --target founder@example.com [--topics "gtm,goal-loop"]
  /notification-router add-log --name local-audit --min-severity warn [--kinds "lite.*"]
  /notification-router remove --name ops-critical
  /notification-router test [--name ops-critical] [--message "..."]
/novel
Serialized novel generation (lite): supports --lang/--format/--genre, series continuation, multi-chapter runs, DOCX/PDF export, and contest mode. Writes to ~/.maria/desktop/artifacts/novel/ by default.
USAGE
/novel "<theme>" [--series <id|@path>] [--new-series] [--lang en|ja] [--format md|txt]
    [--genre <name>] [--chapters N] [--out <dir>] [--pdf] [--docx]
    [--in @requirements.txt] [--contest] [--persona <id>] [--envelope <jsonOr@path>] [--replace]
  
    --pdf               Export as PDF
    --docx              Export as DOCX
    --in @path1,@path2  Contest requirements / reference materials (supports files and folders)
    --contest           Contest mode: generate based on contest requirements from --in
    --persona <id>      Use a registered persona as creative voice (default: auto)
/origin-fetch
Fetch official external API docs and store them under origin (then rebuild origin index + NDC).
USAGE
/origin-fetch "<query>" [--label <name>] [--max-urls 6] [--sample] [--json]
  /origin-fetch --url "https://a.com,https://b.com" [--label <name>] [--sample]
  /origin-fetch --from-file urls.txt [--label <name>] [--sample]
  Notes:
  - Default destination: artifacts/origin/<label>/
  - With --sample: src-lite/origin/<label>/ (publishable)
  - --url: Directly specify URLs (comma-separated). Skips AI discovery.
  - --from-file: Read URLs from a file (one per line). Skips AI discovery.
/paper
Academic paper writing: research, outline, write sections, export PDF+DOCX. Math via LaTeX notation.
EXAMPLES
/paper "Transformer architectures for time-series forecasting"
    /paper --in @notes/draft.md --topic "quantum computing survey" --lang en
    /paper --topic "Self-recognition of autonomous AI" --lang ja
/pdf
Convert Markdown to PDF deterministically via md-to-pdf (stdio mode).
USAGE
/pdf from-md --in <file.md> [--out <file.pdf>] [--basedir <dir>]
  Example: /pdf from-md --in docs/report.md --out artifacts/report.pdf
/persona
Create, manage, and use personas (self or fictional). Personas are referenced by /novel, /manga, /film, /proposal, etc. for creative voice, and by /docs, /spreadsheet for formal auto-fill.
USAGE
/persona register [--self] [--lang ja|en] [--label <name>] [--avatar @img.png]  — Register persona
  /persona register --in @answers.txt [--self] [--lang ja|en] [--avatar @img.png]  — Register from file
  /persona generate "<description>" [--lang ja|en] [--avatar @img.png] [--no-avatar] [--avatar-style realistic|vector|anime|artistic]  — AI-generate a fictional persona (avatar auto-generated unless --no-avatar)
  /persona modify <name> "<instruction>" [--avatar @img.png]  — Modify an existing persona
  /persona modify <name> --auto  — Auto-fill empty fields using AI
  /persona modify <name>  — View current persona in editable format
  /persona list  — List all personas
  /persona list --samples  — List bundled sample personas
  /persona install <name>  — Install a sample or marketplace persona
  /persona share <name>  — Publish a persona to the marketplace
  /persona search "<query>"  — Search marketplace personas
  /persona show <name>  — Show persona details
  /persona remove <name>  — Remove a persona
  /persona set-default <name>  — Set default persona
  /persona memory add <name> "<content>" [--tags t1,t2] [--importance high|medium|low]  — Add memory
  /persona memory list <name>  — List memory entries
  /persona memory show <name>  — Show memory as prompt block
  /persona memory remove <name> <entry-id>  — Remove a memory entry
/phone-deploy
Deploy and manage callAI Cloud Run services.
USAGE
/phone-deploy status [--project <id>] [--region <region>]
  /phone-deploy health [--project <id>]
  /phone-deploy trigger [--target ws|frontend|both] [--branch <branch>] [--project <id>]
  /phone-deploy builds [--limit <n>] [--project <id>]
  /phone-deploy build-log <buildId> [--project <id>]
  /phone-deploy logs [--limit <n>] [--project <id>] [--service <name>]
  /phone-deploy rotate-token [--project <id>]
/phone-dept
Manage CallAI departments.
USAGE
/phone-dept list
  /phone-dept show <departmentId>
  /phone-dept add --id <departmentId> --name <name> [--phone <phone>] [--description <text>]
  /phone-dept update <departmentId> --name <name> [--phone <phone>]
  /phone-dept remove <departmentId>
/phone-dict
Manage CallAI reading dictionary (from/to text replacements).
USAGE
/phone-dict list [--enabled-only]
  /phone-dict add --from <text> --to <text> [--note <text>]
  /phone-dict remove <entryId>
  /phone-dict toggle <entryId>
/phone-hp-import
Manage CallAI HP (homepage) import adapters.
USAGE
/phone-hp-import list
  /phone-hp-import show <adapterId>
  /phone-hp-import test <adapterId> --url <listUrl>
/phone-init
Auto-initialize CallAI tenant from a URL or file. Extracts organization info, departments, and FAQ.
USAGE
/phone-init --url <homepage_url>
  /phone-init --file <path.csv|json>
  /phone-init --url <url> --file <path>
  /phone-init preview --url <url>  (preview only, show extracted data without writing)
/phone-prompt
Manage CallAI prompts and voice profiles.
USAGE
/phone-prompt list
  /phone-prompt show <promptId>
  /phone-prompt set <promptId> --system <text> [--voice-profile <id>]
  /phone-prompt voice-list
  /phone-prompt voice-set <profileId> --voice <text>
/phone-tenant
Manage CallAI tenant configuration.
USAGE
/phone-tenant show
  /phone-tenant set --key <field> --value <value>
  /phone-tenant init --org <organizationName>
  /phone-tenant connect --url <serverUrl> --token <adminToken> [--tenant <tenantId>]
  /phone-tenant disconnect
  /phone-tenant status
/plugins
Claude Code plugin/marketplace compatibility manager (requires --cc).
USAGE
Marketplaces:
    /plugins marketplace add <source> [--scope user|project|local] --cc
    /plugins marketplace list --cc
    /plugins marketplace update <name> --cc
    /plugins marketplace remove <name> [--scope user|project|local] --cc
  
  Plugins:
    /plugins install <plugin>@<marketplace> [--scope user|project|local] --cc
    /plugins enable <plugin>@<marketplace> [--scope user|project|local] --cc
    /plugins disable <plugin>@<marketplace> [--scope user|project|local] --cc
    /plugins uninstall <plugin>@<marketplace> [--scope user|project|local] --cc
    /plugins list [--scope user|project|local] [--with-skills] [--save] [--save-path <path>] [--max-inline-chars <n>] --cc
  
  Marketplace publish (generate marketplace.json):
    /plugins marketplace publish --name <marketplace> --owner-name <name> [--owner-email <email>] --plugins-root <dir> [--out <dir>] --cc
/proposal
Create a customer proposal deck by running /slides with a fixed proposal-oriented prompt and reusable common settings.
EXAMPLES
/proposal set-common-settings --bg #FFFFFF --fg #111111 --accent #22C55E --font "Noto Sans JP" --bg-image @assets/bg.png
    /proposal set-common-settings --title-bg-image @assets/bg-title.png
    /proposal set-common-settings --bg-image @assets/bg.png
    /proposal set-common-settings --title-title-shift "0.0,0.0,0.2,0.0" --body-title-shift "0.0,0.0,0.1,0.0"
    /proposal --target "ACME" --product-info @docs/product.md --customer-memo @docs/acme.md --slides 12 --pdf --pptx
  
  Subcommands:
    set-common-settings         (save/update common bg/fg/accent/font/bg-image for reuse)
  
  Options (generate):
    --target "<name>"           (required)
    --product-info @path1,@path2 (required; supports files and folders; comma-separated)
    --customer-memo @path1,@path2 (optional; supports files and folders; comma-separated)
    --slides N                  (default 10; 1..64)
    --title-slide true|false    (default true)
    --aspect 16:9|4:3           (default 16:9)
    --bg #RRGGBB                (optional; overrides common settings)
    --fg #RRGGBB                (optional; overrides common settings)
    --accent #RRGGBB            (optional; overrides common settings)
    --font "<name>"             (optional; overrides common settings)
    --bg-image @path            (optional; overrides common settings; compose background image png/jpg)
    --title-bg-image @path      (optional; overrides common settings; compose title background image png/jpg)
    --title-title-shift U,R,D,L     (optional; overrides common settings; % of slide; compose only)
    --title-subtitle-shift U,R,D,L  (optional; overrides common settings; % of slide; compose only)
    --body-title-shift U,R,D,L      (optional; overrides common settings; % of slide; compose only)
    --size WxH                  (default 1536x864)
    --format png|jpg|webp       (default png)
    --model <id>                (optional)
    --seed N                    (optional)
    --persona <id>          (optional; creative persona voice)
    --pdf
    --pptx
    --plan-only
/quit
Exit the MARIA LITE session.
/recompile-api
Recompile a registered API catalog from its OpenAPI URL (applies by default; use --plan-only to preview).
USAGE
/recompile-api --service <id> [--update] [--plan-only] [--json]
/record
Control Desktop screenshot and screen recording capture from a slash command.
USAGE
/record [status] [--target screen|camera|avatar] [--json]
  /record start [--target screen|camera|avatar] [--json]
  /record stop [--target screen|camera|avatar] [--json]
  /record screenshot [--target screen|camera|avatar] [--json]
  
  Notes:
  - Run from MARIA Desktop to execute the capture locally.
  - /record start opens the browser/Tauri capture picker when required.
  - target=screen records the desktop/window/tab capture flow.
  - target=camera records the user-facing camera feed.
  - target=avatar records the separate VTuber avatar runtime, not /vui.
  - Screenshots and recordings are saved under desktop/captures/.
/register-api
Register an API (shared catalog in Firebase; secrets stay local). Applies by default; use --plan-only to preview.
USAGE
/register-api --service <name> [--openapi <https_url>]
    [--registrable-domain <example.com>] [--allowed-hosts "*.example.com,api.example.com"]
    [--docs-url <https_url>] [--display-name <name>]
    [--secret <key=value>] [--secrets <json_object>]
    [--render] [--force] [--plan-only] [--json]
/reorganize
Reorganize (lite): restructure local RAG indices into NDC-based shards to keep retrieval scalable as corpora grow.
EXAMPLES
/reorganize --origin
    /reorganize --kp --sample --full
    /reorganize --universe
/research
Research (lite): ask ai-proxy to gather external information and summarize it.
USAGE
/research "<query|url>" [--top-k 5] [--json]
/resource-manager
Manage Google Workspace buildings and calendar resources.
USAGE
/resource-manager buildings
  /resource-manager resources [--building <buildingId>] [--capacity <n>]
  /resource-manager create-building --name "Main Office" [--floors <n>] [--description <text>]
  /resource-manager create-resource --name "Room A" --building <buildingId> [--capacity <n>] [--type CONFERENCE_ROOM] [--floor <floor>] [--description <text>]
  /resource-manager update-building --id <buildingId> [--name <name>] [--floors <n>] [--description <text>]
  /resource-manager update-resource --id <resourceId> [--name <name>] [--building <buildingId>] [--capacity <n>] [--type <type>] [--floor <floor>] [--description <text>]
/restart
Resume a previous run (use --run <runId>).
/review
List, approve, reject, and replay gated executions.
USAGE
/review [list]
  /review approve --run <runId> [--note "..."] [--replay]
  /review reject --run <runId> [--note "..."]
  /review replay --run <runId>
/rollback
Rollback applied changes for dev commands (checkpoint-based).
USAGE
/rollback [--count 1]
  /rollback --run <runId>
  /rollback --list [--json]
  
  Notes:
  - Rollback restores files to the state captured right before an apply run.
  - Use --run to roll back multiple checkpoints at once (newest -> that run, inclusive).
/run-api
Run a registered API action via the fixed executor (no arbitrary curl/code).
USAGE
/run-api --service <id> --action <id> [--input <json>] [--input-file <path>]
    [--json]
EXAMPLES
/run-api --service openai --action openai.list_models --input "{}"
/salesforce
Salesforce GTM wrapper for CRM objects, pipeline orchestration, and reporting APIs.
USAGE
/salesforce setup [--openapi <url>] [--docs-url <url>] [--display-name <name>] [--secret <key=value>] [--secrets <json>] [--plan-only] [--json]
  /salesforce actions [--action <id>] [--json]
  /salesforce run --action <id> [--input <json>] [--input-file <path>] [--json]
  /salesforce connect [--status] [--disconnect] [--action <id>] [--json]
  /salesforce secret [--list] [--required] [--secret <key=value>] [--secrets <json>] [--json]
  /salesforce status [--json]
  Hint: Run `/salesforce setup` first, then `/salesforce connect` or `/salesforce secret --required`.
  Auth: Salesforce commonly uses OAuth; `/salesforce connect` is the preferred flow when available.
/schedule
Local scheduler (lite): run commands periodically via a cross-platform local daemon (win/mac/linux). Supports multiple named schedules.
USAGE
Status:
    /schedule                   # list all schedules
    /schedule status            # list all schedules
    /schedule status --name <id>  # status of a specific schedule
  
  Configure schedule (V2 — named + multi-command):
    /schedule set --name daily-intel --daily-at 07:00 --commands "/daily-report --discord,/daily-news --discord,/daily-insights --discord"
    /schedule set --name weekly-intel --every 7d --commands "/weekly-report --discord"
    /schedule set --name release-health --cron "0 */6 * * *" --commands "/self-improve run --code --auto-fix"
  
  Configure schedule (V1 compat — self-recognition-evolve):
    /schedule set --every 24h --kp --desire [--direction "..."] [--max-topics 5]
    /schedule set --daily-at 03:30 --kp --universe --desire
    - Default behavior also starts (or restarts) the daemon immediately.
  
  Start/stop daemon:
    /schedule start [--name <id>]    # default: start all configured
    /schedule stop [--name <id>]     # default: stop all running
  
  Remove schedule (stop + delete config/state):
    /schedule remove [--name <id>]
  
  Run once (foreground):
    /schedule run [--name <id>]
/self-improve
Autonomous self-improvement loop: reuse recent memory, run self-recognition, dispatch evolve, and optionally attempt low-risk code auto-fix.
USAGE
/self-improve run [--kp] [--universe] [--code] [--desire] [--skills] [--persona] [--direction "..."] [--max-topics 5] [--auto-fix] [--json]
  /self-improve autofix [--direction "..."] [--json]
  /self-improve status [--json]
/self-recognition
Self-recognition (lite): analyze current state and list missing areas for supportive AGI. With --code, include code-quality and UX code audits before proposing gaps.
USAGE
/self-recognition [--kp] [--universe] [--code] [--desire] [--skills] [--persona] [--sample] [--direction "..."] [--max-topics 5] [--json]
/skills
Claude Code-compatible skills runner (reads .claude/skills and ~/.claude/skills).
USAGE
Run a skill:
    /skills --name <skill-name> [args...]
    /skills --name <plugin-name:skill-name> [args...]
/slack-webhook
Configure inbound Slack Events API webhook handling and optional goal-loop triggers.
USAGE
/slack-webhook
  /slack-webhook setup [--signing-secret "..."] [--events "app_mention,message,reaction_added"] [--goal-loop-auto-fix true|false] [--goal-loop-apply true|false] [--goal-loop-max-topics 5]
  /slack-webhook status [--json]
  /slack-webhook disable
/slides
Generate slide images from source material; optionally export PDF/PPTX.
EXAMPLES
/slides --in @docs/brief.md --purpose "Internal briefing" --slides 10
    /slides --purpose "Sales proposal" "Create a 12-slide deck about our Q1 results"
/spreadsheet
Generate spreadsheet data from source material and export to XLSX+PDF.
EXAMPLES
/spreadsheet --in @data/report.md --title "Monthly Report" --xlsx --pdf
    /spreadsheet --title "Budget" "Create a budget spreadsheet for Q1 2026"
/stop
Stop the current run (resume with /restart).
/task-distribution
Task distribution: decompose a goal into tasks, allocate to members with effort estimates and interface contracts.
EXAMPLES
/task-distribution "Build MVP for customer portal" --members "Alice,Bob,Charlie"
    /task-distribution --in @project-brief.md --members "Dev1,Dev2,Designer,PM"
    /task-distribution "Add auth module" --members "Dev1,Dev2" --repo "owner/repo"
    /task-distribution "Refactor payments" --members "A,B" --repo "org/app" --publish-issues
/task-manager
Show system status, running tasks, and universe flow visualization.
EXAMPLES
/task-manager
    /task-manager --json
    /task-manager --flow <universe-name>
/twin
Manage Digital Twin service integrations (email, calendar, chat, dev tools, accounting, etc.).
USAGE
/twin list [--json]                         — Show configured services
  /twin presets [<query>] [--category <cat>] [--json] — Show / search available services
  /twin add <serviceId> [--json]               — Add a service from presets
  /twin remove <serviceId> [--json]            — Remove a service
  /twin status [--json]                        — Check connection status
  /twin setup [<serviceId>] [--json]           — Guided setup wizard
/universe
Universe-code: generate .uni programs (genesis), compile and run (inflate), debug, edit.
SUBCOMMANDS
Genesis
AI generates a .uni program for the given topic (multi-agent team by default).
/universe genesis "topic" --name <name> [--goal "success criteria"] [--code] [--sample] [--json]
Inflate
compile and run a .uni universe (fallback to legacy flow if no .uni).
/universe inflate --name <name> [--sample] [--rerun] [--in <text|@file|@f1,@f2>] [--answer "..."] [--plan-only] [--json]
Debug
parse, compile, validate a .uni file; AI auto-fix on errors.
/universe debug --name <name> [--sample] [--plan-only]
Edit
view/edit the .uni source (opens editor in desktop mode).
/universe edit --name <name> [--sample]
Org chart
print the deterministic org structure (nodes + agents).
/universe org-chart --name <name> [--sample] [--json]
Multiverse
list all known universes (sample + artifacts).
/universe multiverse [--json] [--save] [--no-save] [--save-path artifacts/multiverse.md]
Revelation
append creator-provided knowledge into common-context.
/universe revelation --name <name> --push <text|filepath> [--sample] [--json]
USAGE
Subcommands:
  
  1) Genesis: AI generates a .uni program for the given topic (multi-agent team by default).
     /universe genesis "topic" --name <name> [--goal "success criteria"] [--code] [--sample] [--json]
     - Default: multi-agent team genesis (researcher → builder → reviewer loop).
     - With --code: single-agent genesis (one-shot LLM generation with retry).
     - Writes: universe.uni, universe.uni.ir.json, universe.manifest.json
     - Storage:
       - With --sample: src-lite/universe-code/sample/<name>/
       - Default      : ~/.maria/desktop/universe/<name>/
     - Note: requires authentication for ai-proxy. Run /login first.
  
  2) Inflate: compile and run a .uni universe (fallback to legacy flow if no .uni).
     /universe inflate --name <name> [--sample] [--rerun] [--in <text|@file|@f1,@f2>] [--answer "..."] [--plan-only] [--json]
     /universe inflate "<topic>" [--sample] [--rerun] [--in <text|@file|@f1,@f2>] [--answer "..."] [--plan-only] [--json]  (auto-picks best)
     - Default       : agents execute tools (/code, /research, /deps, etc.).
     - With --plan-only: preview only (no side effects).
     - With --rerun   : discard previous checkpoint and start a fresh run.
     - If blocked, it returns a question. Resume by passing --answer.
  
  3) Debug: parse, compile, validate a .uni file; AI auto-fix on errors.
     /universe debug --name <name> [--sample] [--plan-only]
     - Default         : auto-patches the .uni file.
     - With --plan-only: shows diagnostics and suggested patches only.
  
  4) Edit: view/edit the .uni source (opens editor in desktop mode).
     /universe edit --name <name> [--sample]
  
  5) Org chart: print the deterministic org structure (nodes + agents).
     /universe org-chart --name <name> [--sample] [--json]
  
  6) Multiverse: list all known universes (sample + artifacts).
     /universe multiverse [--json] [--save] [--no-save] [--save-path artifacts/multiverse.md]
  
  7) Revelation: append creator-provided knowledge into common-context.
     /universe revelation --name <name> --push <text|filepath> [--sample] [--json]
EXAMPLES
/universe genesis "Hiring process redesign" --name hiring_demo --sample
    /universe genesis "Hiring process redesign" --name hiring_demo --code --sample
    /universe genesis "Hiring process redesign" --name hiring_demo --concurrency 4
    /universe inflate --name hiring_demo --sample
    /universe debug --name hiring_demo --sample
    /universe edit --name hiring_demo --sample
    /universe multiverse
    /universe revelation --name hiring_demo --sample --push "Always produce a final PDF."
/upgrade
Open plan upgrade flow for Starter, Pro, or Ultra.
USAGE
/upgrade <plan> [--annual] [--monthly] [--no-open] [--json]
/usage
Show current plan usage, remaining quota, and reset date.
USAGE
/usage [--json]
/ux-improvement
Code-based UX audit for a repository: map usability friction, explain causes, and export an improvement report (DOCX+PDF).
EXAMPLES
/ux-improvement "."
    /ux-improvement "../my-web-app" --max-files 160
    /ux-improvement --path "./apps/admin" --repo "owner/repo" --gh-issue
    /ux-improvement --repo "owner/repo" --gh-issue
/version
Show MARIA-LITE version information.
USAGE
/version [--json]
/video
Generate videos via api.maria-code.ai (compatible with /api/v1/video). Saves to outputs/videos by default.
USAGE
/video "prompt" [--duration 8] [--fps 24] [--res 720|1080|1280x720] [--aspect 16:9|9:16] [--format mp4|webm] [--model <id>] [--provider openai|google] [--seed N] [--remix <openai_video_id>] [--ref @image.png] [--out <dir>] [--path <file>] [--plan-only]
/vui
Launch Voice User Interface — speak commands to MARIA with avatar visualization.
USAGE
/vui
  
  Launches the VUI in the Desktop browser. Speak naturally and MARIA
  will analyse your intent, plan commands, and execute them.
  
  Features:
    - WebGL shader orb with voice-reactive animation (default)
    - VRM 3D avatar with FFT lip sync + expressions (Three.js CDN)
    - Live2D 2D avatar with lip sync + emotions (Cubism Framework)
    - Cross-platform mic capture with browser speech recognition or recorder/STT fallback
    - TTS voice reply for chat / command results
    - Intent analysis via command planning with dependency resolution
    - Multi-command chaining and execution
/vup
Bump MARIA-LITE version and write release notes (applies by default; use --plan-only to preview).
USAGE
Apply version bump + README update (default):
    /vup
  
  Preview without writes:
    /vup --plan-only
  
  Commit, tag, and push after /vup:
    /vup push
    (GitHub Actions then handles desktop release + MARIA OS Lite deploy/rollout.)
  
  Set explicit version:
    /vup --version 2.1.5
  
  Set explicit data package version:
    /vup --version 2.1.5 --version-data 1.1.0
/watch
Cross-platform polling file watcher: detect repo changes, publish events, and run follow-up commands.
USAGE
/watch set --name repo-watch --paths "src-lite,maria-code-lp" --interval 30s --commands "/goal-loop run --event-kind fs.changed --auto-fix,/doctor" [--ignore ".git,node_modules,dist,artifacts"] [--debounce 15s]
  /watch status [--name <id>] [--json]
  /watch start [--name <id>]
  /watch stop [--name <id>]
  /watch remove [--name <id>]
/weekly-report
Generate a weekly activity report from GitHub, Google Calendar, and Gmail.
USAGE
/weekly-report [--since <YYYY-MM-DD>] [--days <n>] [--format <markdown|slack|plain|custom>] [--max-repos <n>] [--out <path>] [--json] [--discord [channel]]
  Custom format: paste heading template as prompt text (e.g. "This week's work\nIssues\nNext week's plan")
/work-game
Define business work as a game loop, track actions, generate next actions, and export ready-to-use draft files.
USAGE
/work-game
  /work-game init --name "<team>" [--mission "..."] [--timezone Asia/Tokyo] [--start 09:00] [--end 18:00] [--workdays mon,tue,wed,thu,fri] [--holidays 2026-03-20,2026-05-03] [--discord daily-ops] [--post-holidays] [--lp]
  /work-game add-tool --name "<tool>" [--kind crm|calendar|mail|docs|dev|sns|other] [--vendor "<vendor>"] [--hints "hubspot,crm,meeting.created"] [--notes "..."]
  /work-game add-task --name "<task>" --category mtg|sales|docs|legal|dev|mail|sns|ops|custom [--cadence daily|weekly|adhoc] [--times 1] [--per-week 5] [--duration 45] [--tools "gmail,docs"] [--outputs "docx,mail-draft"] [--window 09:00-11:00] [--priority 1..5] [--prompt "..."] [--notes "..."]
  /work-game log --task <task-id|name> [--tool <tool-id|name>] --action "<text>" [--phase start|progress|complete] [--duration 25] [--artifact docx] [--notes "..."]
  /work-game analyze [--drafts] [--json]
  /work-game drafts [--json]
  /work-game post-discord [--channel <name>] [--force] [--json]
  /work-game export-uni [--out artifacts/maria-lite/work-game/custom.work-game.uni] [--json]
  
  Notes:
  - Drafts write under ~/.maria/desktop/artifacts/maria-lite/work-game/drafts/ by default.
  - Discord posting respects holiday/weekend policy unless --force is used.
  - Use /desktop and open /work-game for the integrated builder UI.