@entelligentsia/forgecli 1.1.0 → 1.1.2

package/dist/forge-payload/init/phases/phase-2/domain-concepts.md

@@ -0,0 +1,26 @@
+<!-- kb-doc-fragment: domain-concepts -->
+# Substance — `business-domain/domain-concepts`
+
+**Output path:** `{kbFolder}/business-domain/domain-concepts.md`
+
+**Topic focus:** the ubiquitous language — the core domain concepts, terms, and
+vocabulary the project uses, with precise definitions. Describe *the words the
+team uses and what they mean*, so a newcomer can read the codebase's language.
+
+**Discovery input to read:** all Phase-1 discovery findings, mined for recurring
+domain vocabulary (names of entities, states, actions, roles). Lean on the
+`database` and `routing` findings for the richest terminology.
+
+**Required output:**
+- Confidence header on line 1.
+- A glossary: each core concept/term with a one-to-two-sentence definition.
+- Note synonyms and any terms with a project-specific (non-obvious) meaning.
+- Concepts only — no field types or route tables (those are other docs).
+
+**Not applicable:** if no distinct domain vocabulary emerges, the topic is *absent*
+— a certainty, not a low-confidence guess. Write a **confident** not-applicable
+stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Domain Concepts` heading and the one-sentence absence statement `No
+specialized domain vocabulary — this project uses only general programming terms.`
+Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/domain-model.md

@@ -0,0 +1,26 @@
+<!-- kb-doc-fragment: domain-model -->
+# Substance — `business-domain/domain-model`
+
+**Output path:** `{kbFolder}/business-domain/domain-model.md`
+
+**Topic focus:** the *conceptual* business model — the real-world entities the
+system represents, their relationships, and the business invariants/rules that
+govern them. This is business meaning, distinct from `entity-model`'s technical
+field inventory.
+
+**Discovery input to read:** the `database` domain findings from Phase 1, read
+through a business lens (what does each stored thing *mean* to the domain).
+
+**Required output:**
+- Confidence header on line 1.
+- Business entities with their conceptual purpose (not field types).
+- Relationships expressed in domain terms (e.g. "a Sprint owns many Tasks").
+- Business invariants and rules (e.g. "a Task cannot be committed before approval").
+- Entity names exactly as in the brief's `## Domain Entities`.
+
+**Not applicable:** if the project has no discernible business domain (e.g. a
+pure tool/utility), the topic is *absent* — a certainty, not a low-confidence
+guess. Write a **confident** not-applicable stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Domain Model` heading and the one-sentence absence statement `No
+business domain — this is a technical utility with no domain entities.` Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/entity-model.md

@@ -0,0 +1,26 @@
+<!-- kb-doc-fragment: entity-model -->
+# Substance — `architecture/entity-model`
+
+**Output path:** `{kbFolder}/architecture/entity-model.md`
+
+**Topic focus:** the *technical* inventory of what is stored — every entity with
+its full field list, types, and structural relationships as they exist in code.
+This is the exhaustive structural catalogue; `domain-model` is its conceptual,
+business-facing counterpart.
+
+**Discovery input to read:** the `database` domain findings from Phase 1 (models,
+schema definitions, type declarations).
+
+**Required output:**
+- Confidence header on line 1.
+- One section per entity with the complete field inventory (name, type, nullability).
+- Structural relationships (FK, embedding, references).
+- Entity names used exactly as they appear in the brief's `## Domain Entities`.
+- Technical inventory only — no business invariants (those live in `domain-model`).
+
+**Not applicable:** if nothing is stored, the topic is *absent* — a certainty,
+not a low-confidence guess. Write a **confident** not-applicable stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Entity Model` heading and the one-sentence absence statement `No
+entities — this project persists no structured data.` Do NOT lower confidence
+— an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/index.md

@@ -0,0 +1,25 @@
+<!-- kb-doc-fragment: index -->
+# Substance — `index` step
+
+**Output paths:** `{kbFolder}/architecture/INDEX.md`,
+`{kbFolder}/business-domain/INDEX.md`, and `{kbFolder}/MASTER_INDEX.md`.
+
+**Topic focus:** build the navigation indexes that link the KB docs already
+written to disk. You run *after* all 10 leaf docs exist. Describe *how to
+navigate the knowledge base*, linking only what is present.
+
+**Discovery input to read:** the filesystem — list the docs actually written
+under `architecture/` and `business-domain/`. Do NOT anticipate files that were
+not produced.
+
+**Required output:**
+1. `architecture/INDEX.md` — list and link every architecture doc on disk.
+2. `business-domain/INDEX.md` — list and link `domain-model.md` and
+   `domain-concepts.md` if present.
+3. `MASTER_INDEX.md` — link both INDEX files and include a `## Domain Entities`
+   section listing discovered entities, one per line.
+- Link **only** docs that exist on disk at write time (AC3).
+- Confidence header on the first line of each written index.
+
+**Not applicable:** never fully skipped — if a section has no docs, write the
+index with an explicit `(none generated)` note rather than a broken link.

package/dist/forge-payload/init/phases/phase-2/processes.md

@@ -0,0 +1,26 @@
+<!-- kb-doc-fragment: processes -->
+# Substance — `architecture/processes`
+
+**Output path:** `{kbFolder}/architecture/processes.md`
+
+**Topic focus:** the runnable processes and build/deploy topology — services,
+daemons, workers, entry points, and the commands that build, start, and run
+them. Describe *what runs* and *how it is built*, not the deployment targets.
+
+**Discovery input to read:** the `processes` domain findings from Phase 1
+(entry points, scripts, service definitions, build commands). Cross-reference
+`stack` for the build tool.
+
+**Required output:**
+- Confidence header on line 1.
+- Each service/process with its entry point and start command.
+- Build pipeline steps in order.
+- Inter-process relationships (who calls/spawns whom), if any.
+- No CI/CD or environment content — that is `deployment`.
+
+**Not applicable:** if the project defines no runnable processes (e.g. a library
+with no service), the topic is *absent* — a certainty, not a low-confidence guess.
+Write a **confident** not-applicable stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Processes` heading and the one-sentence absence statement `No standalone
+processes — this is a library/package consumed by other code.` Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/routing.md

@@ -0,0 +1,25 @@
+<!-- kb-doc-fragment: routing -->
+# Substance — `architecture/routing`
+
+**Output path:** `{kbFolder}/architecture/routing.md`
+
+**Topic focus:** the API/route surface — endpoints or command surface, route
+groups, HTTP methods (or CLI verbs), and the auth/authorization strategy that
+gates them. Describe *how the system is entered*, not what it stores.
+
+**Discovery input to read:** the `routing` domain findings from Phase 1 (route
+definitions, controllers, handlers, middleware, auth guards).
+
+**Required output:**
+- Confidence header on line 1.
+- Route/endpoint groups with methods and paths (or CLI command surface).
+- Auth strategy and where it is enforced.
+- Middleware / interceptors in the request path.
+- No persistence or deployment content.
+
+**Not applicable:** if the project exposes no routes or command surface (e.g. a
+pure library), the topic is *absent* — a certainty, not a low-confidence guess.
+Write a **confident** not-applicable stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Routing` heading and the one-sentence absence statement `No routing
+surface — this project exposes no HTTP routes or command entry points.` Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/stack-checklist.md

@@ -0,0 +1,25 @@
+<!-- kb-doc-fragment: stack-checklist -->
+# Substance — `architecture/stack-checklist`
+
+**Output path:** `{kbFolder}/architecture/stack-checklist.md`
+
+**Topic focus:** an actionable review checklist derived from the stack and
+testing facts — the concrete gates a change in this codebase must pass (syntax
+check, lint, test, build, type-check commands). Describe *what to verify before
+a change is considered done*.
+
+**Discovery input to read:** the `stack` and `testing` domain findings from
+Phase 1 (build tool, linter, type checker, test runner, their commands).
+
+**Required output:**
+- Confidence header on line 1.
+- A checklist of verification steps, each with its exact command.
+- Ordered from cheapest (syntax) to most expensive (full suite/build).
+- Derived strictly from stack + testing facts — no invented tooling.
+
+**Not applicable:** if no verification tooling exists, the topic is *absent* — a
+certainty, not a low-confidence guess. Write a **confident** not-applicable stub:
+line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Stack Checklist` heading and the one-sentence absence statement `No
+verification tooling detected — no automated checklist can be derived.` Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/stack.md

@@ -0,0 +1,27 @@
+<!-- kb-doc-fragment: stack -->
+# Substance — `architecture/stack`
+
+**Output path:** `{kbFolder}/architecture/stack.md`
+
+**Topic focus:** the technology stack — languages and their versions, frameworks
+and major libraries, the runtime/interpreter, package manager, and the build
+toolchain. Describe *what the project is built with*, not how it is deployed.
+
+**Discovery input to read:** the `stack` domain findings from Phase 1
+(languages, dependency manifests, lockfiles, runtime versions). Cross-reference
+the `testing` findings only for the test-runner dependency.
+
+**Required output:**
+- Confidence header on line 1.
+- A languages/versions table or list (each fact `[?]`-marked if not directly observed).
+- Frameworks & major libraries with their roles.
+- Runtime + package manager + build tool.
+- No deployment, routing, or schema content — those are other docs.
+
+**Not applicable:** if the project has no identifiable stack (e.g. an empty or
+docs-only repo), the topic is *absent* — a certainty, not a low-confidence guess.
+Write a **confident** not-applicable stub: line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Stack` heading and the one-sentence absence statement `No application
+stack detected — this repository contains no build- or runtime-bearing sources.`
+Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2/testing.md

@@ -0,0 +1,26 @@
+<!-- kb-doc-fragment: testing -->
+# Substance — `architecture/testing`
+
+**Output path:** `{kbFolder}/architecture/testing.md`
+
+**Topic focus:** the test strategy — frameworks and runners, the exact command
+to run the suite, test layout (unit/integration/e2e), and coverage tooling.
+Describe *how the project is verified*.
+
+**Discovery input to read:** the `testing` domain findings from Phase 1 (test
+directories, test frameworks, runner config, coverage config). Cross-reference
+`stack` for the runner dependency version.
+
+**Required output:**
+- Confidence header on line 1.
+- Test framework(s) and the resolved run command (from `commands.test`).
+- Test layout and naming conventions.
+- Coverage tool and threshold, if any.
+- No CI wiring — that is `deployment`.
+
+**Not applicable:** if the project has no tests, the topic is *absent* — a
+certainty, not a low-confidence guess. Write a **confident** not-applicable stub:
+line 1 is
+`<!-- AUTO-GENERATED by /forge:init — confidence: 100% — status: not-applicable -->`,
+then the `# Testing` heading and the one-sentence absence statement `No test suite
+detected — this project currently has no automated tests.` Do NOT lower confidence — an absent topic is known with certainty.

package/dist/forge-payload/init/phases/phase-2-discover.md

@@ -1,6 +1,6 @@
 # Phase 2 — Discover
 
-**Deliverable:** 7 KB architecture docs + 3 index files + `project-context.json` + calibration baseline.
+**Deliverable:** 10 KB docs (8 architecture + 2 business-domain) + 3 index files + `project-context.json` + calibration baseline.
 
 Set `$FORGE_ROOT` and resolve `$KB_PATH` from `.forge/config.json`:
 
@@ -49,31 +49,37 @@ touch "{KB_PATH}/sprints/.gitkeep" "{KB_PATH}/bugs/.gitkeep" \
 
 Read `$FORGE_ROOT/init/generation/generate-kb-doc.md` once (the per-subagent rulebook).
 
-Generate all 7 knowledge-base documents. For each document, analyse the
+Generate all 10 knowledge-base documents. For each document, analyse the
 project codebase for that topic and write to its output path. After writing
 each document, read it back and verify the confidence header is present
 (`<!-- AUTO-GENERATED — confidence: NN% -->`).
 
+Output paths below are the canonical `{KB_PATH}/{docId}.md` targets for the
+shared 10-doc contract (verify-phase.cjs `--phase 2` checks exactly these).
+
 | Document | Output path | Focus |
 |----------|-------------|-------|
 | stack.md | `{KB_PATH}/architecture/stack.md` | Languages, frameworks, runtime, versions |
 | processes.md | `{KB_PATH}/architecture/processes.md` | Services, build/deploy topology |
-| database.md | `{KB_PATH}/architecture/database.md` | Entities, relationships, field types |
 | routing.md | `{KB_PATH}/architecture/routing.md` | API surface, route groups, auth strategy |
+| database.md | `{KB_PATH}/architecture/database.md` | Entities, relationships, field types |
+| testing.md | `{KB_PATH}/architecture/testing.md` | Test frameworks, run commands, coverage |
 | deployment.md | `{KB_PATH}/architecture/deployment.md` | Environments, CI/CD, infra targets |
-| entity-model.md | `{KB_PATH}/business-domain/entity-model.md` | Full entity inventory with fields |
-| stack-checklist.md | `{KB_PATH}/stack-checklist.md` | Review checklist items from stack + testing |
+| entity-model.md | `{KB_PATH}/architecture/entity-model.md` | Full entity inventory with fields |
+| stack-checklist.md | `{KB_PATH}/architecture/stack-checklist.md` | Review checklist items from stack + testing |
+| domain-model.md | `{KB_PATH}/business-domain/domain-model.md` | Business entities, relationships, invariants |
+| domain-concepts.md | `{KB_PATH}/business-domain/domain-concepts.md` | Ubiquitous language, core domain concepts |
 
-You may spawn all 7 as parallel subagents in a single Agent tool message for speed.
-Wait for all 7 to return. Retry any that returned FAILED: once.
+You may spawn all 10 as parallel subagents in a single Agent tool message for speed.
+Wait for all 10 to return. Retry any that returned FAILED: once.
 Any still failing after one retry: halt and surface the id list.
 
 ### Step 4 — Create index files (sequential)
 
 After all leaf docs are written:
 
-1. **`{KB_PATH}/architecture/INDEX.md`** — list and link to the 5 architecture docs
-2. **`{KB_PATH}/business-domain/INDEX.md`** — list and link to entity-model.md
+1. **`{KB_PATH}/architecture/INDEX.md`** — list and link to the 8 architecture docs
+2. **`{KB_PATH}/business-domain/INDEX.md`** — list and link to domain-model.md and domain-concepts.md
 3. **`{KB_PATH}/MASTER_INDEX.md`** — scaffold linking both INDEX files; include
    `## Domain Entities` section listing discovered entities (one per line)
 

package/dist/forge-payload/payload-manifest.json

@@ -210,10 +210,10 @@
       "install": ".forge/init/phases/",
       "owner": "forge-scaffold",
       "select": {
-        "recursive": false,
+        "recursive": true,
         "ext": [".md"]
       },
-      "note": "Init rulebook phase prompts read by run-phases.ts / wfl-init.js (build-payload Pass-2 2e-pre)."
+      "note": "Init rulebook phase prompts read by run-phases.ts / wfl-init.js (build-payload Pass-2 2e-pre). Recursive so the phase-2/ per-step substance fragments (index.md, context.md, and the 10 KB-doc fragments) vendor into .forge/init/phases/phase-2/ — wfl-init.js references them literally (FORGE-S35-T06 / Slice 5)."
     },
     {
       "source": "init/discovery",

package/dist/forge-payload/tools/seed-store.cjs

@@ -215,6 +215,20 @@ try {
     }
   }
 
+  // Always lay down the store skeleton — a COLLATION_STATE.json baseline is
+  // written even when zero entities are seeded, so a fresh `/forge:init`
+  // leaves a consistent `.forge/store/` (dir + watermark) instead of nothing.
+  // The next real `/forge:collate` overwrites this with live counts.
+  if (!DRY_RUN) {
+    Store.writeCollationState({
+      collatedAt: new Date().toISOString(),
+      featureCount: 0,
+      sprintCount,
+      taskCount,
+      bugCount,
+    });
+  }
+
   const prefix_ = DRY_RUN ? '[dry-run] ' : '';
   console.log(`${prefix_}Seeded: ${sprintCount} sprint(s), ${taskCount} task(s), ${bugCount} bug(s)`);
 } catch (e) {

package/dist/forge-payload/tools/verify-phase.cjs

@@ -166,19 +166,76 @@ function verifyPhase1Foundation(cwd) {
   };
 }
 
-// ── Phase 2: KB architecture docs verification ────────────────────────────────
+// ── Phase 2: KB documents verification ────────────────────────────────────────
 
-const ARCH_DOCS = ['stack', 'processes', 'database', 'routing', 'deployment', 'entity-model', 'stack-checklist'];
+// The canonical 10-doc KB contract (FORGE-S35-T01). docIds carry their KB folder
+// prefix (architecture/ or business-domain/); the gate resolves each as
+// {kbPath}/{docId}.md — no hardcoded architecture/ prefix.
+//
+// DRIFT-GUARD: this list MUST stay byte-identical (same items, same order) to
+// KB_DOC_IDS in init/base-pack/workflows-js/wfl-init.js and in forge-cli's
+// run-init-types.ts. Asserted by tools/__tests__/kb-doc-contract.test.cjs.
+const ARCH_DOCS = [
+  'architecture/stack',
+  'architecture/processes',
+  'architecture/routing',
+  'architecture/database',
+  'architecture/testing',
+  'architecture/deployment',
+  'architecture/entity-model',
+  'architecture/stack-checklist',
+  'business-domain/domain-model',
+  'business-domain/domain-concepts',
+];
+
+// A doc satisfies the gate when it exists AND carries a confidence-header token.
+// Keyed on the `confidence: N%` token only (real digit required, case-
+// insensitive) so BOTH in-tree header forms pass:
+//   <!-- AUTO-GENERATED by /forge:init — confidence: 85% -->  (generate-kb-doc.md Rule 2)
+//   <!-- AUTO-GENERATED — confidence: 85% -->                  (phase-2-discover.md Step 3)
+const CONFIDENCE_HEADER = /confidence:\s*\d+%/i;
+
+// A not-applicable topic is a certainty, not a guess: an absent-topic stub is a
+// confident (100%) doc carrying this marker (FORGE-S35-T04 / Slice 3). The marker
+// may live inside the AUTO-GENERATED header comment or in the body. A doc bearing
+// it satisfies the substantive-OR-not-applicable gate even with no other body.
+const STATUS_NOT_APPLICABLE = /status:\s*not-applicable/i;
+
+// Strip HTML comments (the AUTO-GENERATED confidence/status header lives in one)
+// so a doc that is nothing but its header is recognised as empty.
+const HTML_COMMENT = /<!--[\s\S]*?-->/g;
 
 function verifyPhase2(cwd, kbPath) {
   const missing = [];
   const checked = [];
 
   for (const doc of ARCH_DOCS) {
-    const rel = path.join(kbPath, 'architecture', `${doc}.md`);
+    const rel = path.join(kbPath, `${doc}.md`);
     checked.push(rel);
-    if (!fs.existsSync(path.join(cwd, rel))) {
+    const abs = path.join(cwd, rel);
+    if (!fs.existsSync(abs)) {
       missing.push(rel);
+      continue;
+    }
+    let content = '';
+    try {
+      content = fs.readFileSync(abs, 'utf8');
+    } catch {
+      content = '';
+    }
+    if (!CONFIDENCE_HEADER.test(content)) {
+      missing.push(`${rel} (no confidence header)`);
+      continue;
+    }
+    // Substantive-OR-not-applicable gate (FORGE-S35-T04): a doc passes iff it
+    // carries a `status: not-applicable` marker OR has non-empty body content
+    // after the header. Header-only/empty docs are rejected. Completeness stays
+    // ADVISORY — no confidence-percentage threshold, no entity-count bar — so an
+    // entity-less/library project passes with confident not-applicable stubs.
+    const hasNotApplicable = STATUS_NOT_APPLICABLE.test(content);
+    const body = content.replace(HTML_COMMENT, '').trim();
+    if (!hasNotApplicable && body.length === 0) {
+      missing.push(`${rel} (empty — no substance and no status: not-applicable marker)`);
     }
   }
 

package/node_modules/@earendil-works/pi-coding-agent/npm-shrinkwrap.json

@@ -474,8 +474,8 @@
 			}
 		},
 		"node_modules/@earendil-works/pi-agent-core": {
-			"version": "0.80.2-forge.1",
-			"resolved": "https://registry.npmjs.org/@earendil-works/pi-agent-core/-/pi-agent-core-0.80.2-forge.1.tgz",
+			"version": "0.80.2",
+			"resolved": "https://registry.npmjs.org/@earendil-works/pi-agent-core/-/pi-agent-core-0.80.2.tgz",
 			"license": "MIT",
 			"dependencies": {
 				"@earendil-works/pi-ai": "^0.80.2",
@@ -488,8 +488,8 @@
 			}
 		},
 		"node_modules/@earendil-works/pi-ai": {
-			"version": "0.80.2-forge.1",
-			"resolved": "https://registry.npmjs.org/@earendil-works/pi-ai/-/pi-ai-0.80.2-forge.1.tgz",
+			"version": "0.80.2",
+			"resolved": "https://registry.npmjs.org/@earendil-works/pi-ai/-/pi-ai-0.80.2.tgz",
 			"license": "MIT",
 			"dependencies": {
 				"@anthropic-ai/sdk": "0.91.1",
@@ -512,8 +512,8 @@
 			}
 		},
 		"node_modules/@earendil-works/pi-tui": {
-			"version": "0.80.2-forge.1",
-			"resolved": "https://registry.npmjs.org/@earendil-works/pi-tui/-/pi-tui-0.80.2-forge.1.tgz",
+			"version": "0.80.2",
+			"resolved": "https://registry.npmjs.org/@earendil-works/pi-tui/-/pi-tui-0.80.2.tgz",
 			"license": "MIT",
 			"dependencies": {
 				"get-east-asian-width": "1.6.0",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@entelligentsia/forgecli",
-  "version": "1.1.0",
+  "version": "1.1.2",
   "description": "Forge SDLC ported onto @earendil-works/pi-coding-agent — production launcher with three bin aliases (forge/forgecli/4ge). Bundles a curated fork of pi-coding-agent vendored under earendil-works names.",
   "license": "MIT",
   "author": "Entelligentsia",
@@ -49,7 +49,7 @@
     ]
   },
   "forge": {
-    "bundledVersion": "1.4.4",
+    "bundledVersion": "1.6.10",
     "forgeRoot": "../forge/forge"
   },
   "scripts": {