substrate-ai 0.1.24 → 0.1.26

package/dist/cli/templates/statusline.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+# Substrate AI — persistent status line
+# Receives JSON on stdin with session metadata
+
+input=$(cat)
+
+MODEL=$(echo "$input" | jq -r '.model.display_name // "Claude"' 2>/dev/null)
+PCT=$(echo "$input" | jq -r '.context_window.used_percentage // 0' 2>/dev/null | cut -d. -f1)
+COST=$(echo "$input" | jq -r '.session.cost // "0.00"' 2>/dev/null)
+BRANCH=$(echo "$input" | jq -r '.git.branch // ""' 2>/dev/null)
+
+BRANCH_PART=""
+if [ -n "$BRANCH" ]; then
+  BRANCH_PART=" | $BRANCH"
+fi
+
+echo "⚡ substrate-ai | $MODEL | ctx ${PCT}% | \$${COST}${BRANCH_PART}"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "substrate-ai",
-  "version": "0.1.24",
+  "version": "0.1.26",
   "description": "Substrate — multi-agent orchestration daemon for AI coding agents",
   "type": "module",
   "license": "MIT",

package/packs/bmad/manifest.yaml

@@ -58,6 +58,8 @@ phases:
             source: "step:planning-step-2-frs"
           - placeholder: technology_constraints
             source: "decision:analysis.technology-constraints"
+          - placeholder: concept
+            source: "param:concept"
         critique: true
   - name: ux-design
     description: UX discovery, design system, and user journey mapping (optional — runs when uxDesign is true)

package/packs/bmad/prompts/analysis-step-2-scope.md

@@ -36,6 +36,7 @@ Building on the vision analysis above, define the **scope**: core features, succ
    - Extract explicit technology preferences or exclusions stated in the concept
    - Cloud platform choices (e.g., "GCP", "AWS"), programming language mandates (e.g., "Kotlin/JVM", "Node.js excluded"), framework preferences, infrastructure choices
    - If the concept has a "Technology Constraints" section, extract ALL items from it into this field
+   - **Preserve the FULL scope of the user's restriction** — if the user says a technology is "not the right choice" for the backend, do NOT narrow it to a specific subsystem (e.g., "excluded from critical path only"). The exclusion applies to ALL backend services unless the user explicitly limits the scope.
    - Include ONLY preferences explicitly stated by the user — do not infer or add your own
    - If none are stated in the concept, emit an empty array
 

package/packs/bmad/prompts/planning-step-3-nfrs.md

@@ -14,6 +14,9 @@
 ### Technology Constraints (from Analysis Phase)
 {{technology_constraints}}
 
+### Original Concept (user's exact words)
+{{concept}}
+
 ---
 
 ## Mission
@@ -32,7 +35,9 @@ Complete the PRD by defining **non-functional requirements**, **tech stack**, **
    - Key-value pairs mapping technology concerns to specific choices
    - Use real, current technologies — do not fabricate frameworks
    - Cover at minimum: language, framework, database, testing
-   - **MUST honor stated technology constraints** — if the analysis specifies a cloud platform, language, or framework preference, use it. Do not substitute alternatives unless the constraint is technically impossible for the requirements.
+   - **MUST honor stated technology constraints** — if the analysis or original concept specifies a cloud platform, language, or framework preference, use it. Do not substitute alternatives unless the constraint is technically impossible for the requirements.
+   - **When the original concept and the analysis disagree on the scope of a technology restriction, the original concept wins.** For example, if the concept says "JavaScript is not the right choice for backend" but the analysis narrowed it to "excluded from critical path only," honor the concept's broader restriction.
+   - If a technology constraint discourages or excludes a language or runtime (e.g., "JavaScript/Node.js is almost certainly not the right choice"), do NOT use that language or any of its variants (including TypeScript on Node.js) for ANY backend service. This is a blanket backend exclusion, not a path-specific one.
    - If you must deviate from a stated constraint, explicitly note the deviation and rationale
 
 3. **Build the domain model:**
@@ -42,6 +47,12 @@ Complete the PRD by defining **non-functional requirements**, **tech stack**, **
 
 4. **Define out-of-scope items** to prevent scope creep — what this product explicitly does NOT do.
 
+## Pre-Output Checklist
+
+Before emitting your output, verify:
+- [ ] The `language` field in `tech_stack` matches the technology constraints. If a constraint says to evaluate Go, Kotlin/JVM, or Rust (and NOT JavaScript/Node.js), then `language` MUST be one of those — not TypeScript, not JavaScript, not Node.js.
+- [ ] The example below is a FORMAT example only — do not copy its technology choices. Choose technologies based on the constraints above.
+
 ## Output Contract
 
 Emit ONLY this YAML block as your final output — no other text.
@@ -56,10 +67,10 @@ non_functional_requirements:
   - description: "All user data encrypted at rest using AES-256"
     category: "security"
 tech_stack:
-  language: "TypeScript"
-  framework: "Node.js CLI with Commander"
-  database: "SQLite via better-sqlite3"
-  testing: "Vitest"
+  language: "Kotlin 2.0 on JVM 21"
+  framework: "Ktor with coroutines"
+  database: "PostgreSQL 16 via Exposed ORM"
+  testing: "Kotest with Testcontainers"
 domain_model:
   Habit:
     attributes: ["name", "frequency", "created_at"]