Natural Language

edikt teaches Claude to respond to how you naturally talk — no need to remember slash commands.

After /edikt:init, your CLAUDE.md includes a trigger table. Claude reads it at session start and knows what to do when you ask everyday questions.

How it works in practice

You're mid-session, deep in a feature. You want to know where things stand:

You: what's our status?

Claude: (runs /edikt:status)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 EDIKT STATUS — Orders API
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

 ACTIVE PLAN
 ───────────
 PLAN-bulk-orders  ·  2/4 phases (50%)

 | Phase | Title         | Status      |
 |-------|---------------|-------------|
 | 1     | Schema        | done        |
 | 2     | Domain model  | done        |
 | 3     | HTTP handler  | in-progress |
 | 4     | Tests         | -           |

 WHAT'S NEXT
 ───────────
 Phase 3 — HTTP handler
   - Wire up POST /orders/bulk in Chi router
   - Validate with domain service
   - Return 207 multi-status response

You didn't type /edikt:status. You just asked. Claude knew.

---

You come back after the weekend and want Claude to catch up:

You: remind yourself about this project

Claude: (runs /edikt:context)

Loaded context for Orders API:

  Project:   Go REST API, DDD, Chi, PostgreSQL
  Plan:      PLAN-bulk-orders (Phase 3 in progress)
  Rules:     code-quality, testing, security, error-handling, go, chi
  Decisions: 3 ADRs (hexagonal arch, error strategy, auth pattern)
  Product:   spec + 2 PRDs

Ready. What are we working on?

---

You want to start planning a new feature:

You: let's plan the export to CSV feature

Claude: (runs /edikt:plan) Great, let me ask a few questions to scope this out...

---

Full trigger list

You say...	Claude runs
"what's our status?"	/edikt:status
"where are we?"	/edikt:status
"what's next?"	/edikt:status
"what should we work on?"	/edikt:status
"remind yourself"	/edikt:context
"load context"	/edikt:context
"what's this project?"	/edikt:context
"let's plan X"	/edikt:plan
"create a plan for X"	/edikt:plan
"save this decision"	/edikt:adr
"record this choice"	/edikt:adr
"capture that"	/edikt:adr
"that's a hard rule"	/edikt:invariant
"never do X"	/edikt:invariant
"write a PRD for X"	/edikt:prd
"document this feature"	/edikt:prd
"what agents do we have?"	/edikt:agents
"add the security agent"	/edikt:agents add security
"setup Linear"	/edikt:mcp add linear
"validate my environment"	/edikt:team setup

These are defined in .claude/CLAUDE.md — you can add your own triggers by editing that file.

Adding your own triggers

Open .claude/CLAUDE.md and add rows to the trigger table:

| "run the tests" | `make test` |
| "deploy to staging" | `make deploy-staging` |
| "what broke?" | `/edikt:status` then check recent git log |

Any instruction you find yourself repeating to Claude belongs here.