Slash commands
All commands are available through the /ritual entry point inside any supported coding agent. Each command auto-detects light vs cloud mode.
Quick reference
| Command | Light mode | Cloud mode | Time |
|---|---|---|---|
/ritual explore <feature> | Scan codebase → classified questions → RITUAL.md | Create exploration in app.ritual.work | 2–10 min |
/ritual run | Answer questions → recommendations → brief | Run AI pipeline → recommendations → brief | 3–10 min |
/ritual brief | Read RITUAL.md → build-brief.md | Read exploration → Build Brief + Review Resolver | 2–5 min |
/ritual gate | Check code against build-brief.md | Same (always local) | 1–3 min |
/ritual build <feature> | Explore → run → brief → implement → PR | Full pipeline → implement → PR | 30–60 min |
/ritual spec <feature> | Explore → run → brief (no implementation) | Full pipeline → plan only | 15–30 min |
/ritual gherkin | Generate from RITUAL.md | Generate from exploration | 5–10 min |
/ritual questions | — | Generate additional discovery questions | 2–5 min |
/ritual source <context> | Enrich → local .md file | Enrich → add to exploration knowledge sources | 2–5 min |
/ritual answer | Codebase analysis → RITUAL.md | Codebase analysis → upload as knowledge source | 3–10 min |
/ritual context [feature] | — | Load exploration context into the session | 1 min |
/ritual ready | — | Pre-build readiness check | instant |
/ritual conflict | — | Cross-exploration anti-goal scan | 1–5 min |
/ritual lineage <file> | — | Find which explorations shaped a file | 1–2 min |
/ritual drift | — | Check code alignment with recommendations | 3–10 min |
/ritual recap | — | Stakeholder-friendly summary | 1–2 min |
/ritual help | Show commands | Show commands | instant |
The core loop
/ritual explore
Discover what to build before coding.
/ritual explore Add multi-tenant RBAC with team managementScans the codebase, generates classified discovery questions (🔴 blocking, 🟡 uncertain, 🟢 answerable), and saves them to RITUAL.md (light) or an exploration in app.ritual.work (cloud).
Run it before starting any feature. The 🔴 blocking questions are the ones that, if skipped, get flagged in code review.
/ritual run
Answer the questions and generate recommendations.
/ritual runLight mode reads RITUAL.md, answers each question from codebase analysis, and writes build-brief.md. Cloud mode runs the agentic pipeline with knowledge-source enrichment, accepts recommendations, generates requirement packages, and writes the Build Brief.
/ritual brief
Generate a prioritized implementation guide.
/ritual briefLeads with “These Will Block Review If Missing” — review-blocking requirements covering both technical and product completeness. Usually auto-generated by /ritual run; run standalone to regenerate after updating answers.
/ritual gate
Pre-PR adherence check. Always local.
/ritual gateReads the Build Brief’s review-blocking items, scans the codebase for evidence, and reports ✅ Built, ⚠️ Built with limitation, 📋 Pre-declared deferral, or ❌ Not built. PASS if zero ❌ items remain.
/ritual build
Full end-to-end workflow — feature description in, implemented code and drafted PR description out.
/ritual build Add multi-tenant RBAC with team managementRuns reconnaissance, creates or reuses an exploration, generates considerations and questions, runs the AI pipeline, produces a Build Brief, enters plan mode, implements, and drafts .github/PR_DESCRIPTION.md ready for gh pr create --body-file.
Related
- CLI — install and modes.
- MCP server — what Cloud mode connects to.