@entelligentsia/forgecli 1.0.10 → 1.0.14

package/dist/forge-payload/commands/update.md

@@ -242,7 +242,7 @@ Now evaluate — **stop at the first matching row and follow only that row's act
 
 | # | Condition | Action |
 |---|-----------|--------|
-| 1 | `REMOTE_VERSION` == `LOCAL_VERSION` and `LOCAL_VERSION` == baseline | Print "Forge {LOCAL_VERSION} — up to date. No pending migrations." Then execute **Step 4 config refresh** (paths.forgeRoot, paths.forgeRef, backfill) and proceed to **Step 5**. |
+| 1 | `REMOTE_VERSION` == `LOCAL_VERSION` and `LOCAL_VERSION` == baseline | Print "Forge {LOCAL_VERSION} — up to date. No pending migrations." Then execute **Step 4 config refresh** (paths.forgeRef, backfill) and proceed to **Step 5**. |
 | 2 | `REMOTE_VERSION` == `LOCAL_VERSION` and `LOCAL_VERSION` != baseline | Jump to **Step 2B** (project migration — no install needed). |
 | 3 | `IS_CANARY` is true | Jump to **Step 2B** (canary — no install needed). |
 | 4 | `LOCAL_VERSION` > `REMOTE_VERSION` | Print "Local version ({LOCAL_VERSION}) is ahead of the release channel ({REMOTE_VERSION}). No install needed — applying any pending project migrations." then jump to **Step 2B**. |
@@ -459,27 +459,17 @@ path) — skip the re-derivation and keep the original value.
 node "$FORGE_ROOT/tools/banners.cjs" --phase 4 7 "Apply migrations" forge
 ```
 
-> **Sequencing note:** `paths.forgeRoot` is written at the very start of Step 4,
-> before any migration targets or regeneration commands execute. This ensures all
-> subsequent tool invocations in Step 4 (including `build-init-context.cjs` called
-> by regeneration sub-steps) use the current, correct plugin path.
->
-> **Config refresh always runs.** The Step 4 header section (forgeRoot, forgeRef,
-> backfill) executes regardless of whether migrations are pending — even Row 1
+> **Config refresh always runs.** The Step 4 header section (forgeRef, backfill)
+> executes regardless of whether migrations are pending — even Row 1
 > ("up to date") proceeds through this section before skipping to Step 5. The
 > "skip to Step 5" directive skips only the migration chain walk and regeneration.
 > Missing config fields can accumulate across version boundaries; backfill ensures
 > the config stays structurally complete after every `/forge:update` invocation.
 
-**Refresh `paths.forgeRoot` before applying migrations:**
-
-```sh
-node "$FORGE_ROOT/tools/manage-config.cjs" set paths.forgeRoot "$FORGE_ROOT"
-```
-
-**Write `paths.forgeRef` (FR-010):** Also write the installed plugin version
+**Write `paths.forgeRef` (FR-010):** Write the installed plugin version
 as `paths.forgeRef` to config. This makes the config portable across machines —
-`forgeRef` is a version string rather than an absolute path:
+`forgeRef` is a version string rather than an absolute path, and is used by
+`forge-preflight.cjs` to resolve the plugin root via cache lookup:
 
 ```sh
 LOCAL_VERSION=$(node -e "console.log(require('$FORGE_ROOT/.claude-plugin/plugin.json').version)")
@@ -491,7 +481,7 @@ required or recommended fields may have been added since the project was last
 initialized. `manage-config backfill` reads the config schema, compares it
 against the current `.forge/config.json`, and writes defaults for any missing
 fields with schema-defined defaults. It also stamps the top-level `version`
-field from the bundled plugin version. Run after setting forgeRoot/forgeRef:
+field from the bundled plugin version. Run after setting forgeRef:
 
 ```sh
 node "$FORGE_ROOT/tools/manage-config.cjs" backfill --forge-root "$FORGE_ROOT"
@@ -576,12 +566,14 @@ Execute regeneration targets in this order:
 | 5 | `commands` | Must run after `workflows` | — |
 | 6 | `knowledge-base` sub-targets | — (independent) | — |
 | 7 | `schemas` | — (independent) | **Run schema refresh inline** (see Schema Refresh below). Emit: `〇 Refreshing schemas…` |
+| 8 | `workflows-js` | — (independent) | Deterministic verbatim copy of `.claude/workflows/*.js` from base-pack via `/forge:rebuild workflows-js` (no LLM, no placeholder substitution). |
 
 > **Known special targets — note for migration authors:** `hooks` and `schemas` are
 > special-cased here. Future `migrations.json` entries should only use recognised
 > target names; using unknown bare-category targets will produce a warning and be
 > skipped. The recognised targets are: `hooks`, `tools`, `workflows`, `templates`,
-> `personas`, `commands`, `knowledge-base`, `skills`, `schemas`.
+> `personas`, `commands`, `knowledge-base`, `skills`, `schemas`, `workflows-js`.
+> `workflows-js` accepts granular sub-targets (e.g. `workflows-js:wfl-run-task`).
 
 `commands` depends on `workflows` because command wrappers reference workflow
 filenames. All other targets are independent and could run in parallel, but
@@ -666,16 +658,16 @@ reading and following `$FORGE_ROOT/commands/regenerate.md`:
 **Category-to-command mapping:** most categories are handled by
 `/forge:rebuild`, but the `tools` and `schemas` categories are special.
 
-When `tools` appears in the aggregated result, run the schema refresh inline
-(see **Schema Refresh** section below) instead of invoking `/forge:rebuild tools`.
-Tools ship with the plugin and are invoked directly via `$FORGE_ROOT/tools/`.
+When `tools` appears in the aggregated result, invoke `/forge:rebuild tools`
+to re-copy the current plugin's tools closure into `.forge/tools/`. This is
+the actual re-vendor step — do NOT run schema refresh inline instead.
 
 When `schemas` appears in the aggregated result, run the schema refresh inline
-(same **Schema Refresh** section). Do NOT delegate to the removed `/forge:update-tools` command.
+(see **Schema Refresh** section below). Do NOT delegate to the removed `/forge:update-tools` command.
 
 ### Schema Refresh
 
-When the migration chain includes a `schemas` or `tools` target, refresh schemas inline:
+When the migration chain includes a `schemas` target, refresh schemas inline:
 
 ```sh
 mkdir -p .forge/schemas
@@ -1004,8 +996,18 @@ engineer_update_implementation.md → renamed to update_implementation.md
 engineer_fix_bug.md          → renamed to fix_bug.md
 supervisor_review_plan.md    → renamed to review_plan.md
 supervisor_review_implementation.md → renamed to review_code.md
+orchestrate_task.md          → retired (v1.2.0); orchestration runs through .claude/workflows/wfl-run-task.js
+run_sprint.md                → retired (v1.2.0); orchestration runs through .claude/workflows/wfl-run-sprint.js
+fix_bug.md                   → retired (v1.2.0); orchestration runs through .claude/workflows/wfl-fix-bug.js
 ```
 
+> **Note:** `orchestrate_task.md` / `run_sprint.md` / `fix_bug.md` are LLM
+> orchestration prose retired in v1.2.0 — they are no longer generated, and the
+> deterministic JS drivers in `.claude/workflows/wfl-*.js` are the only
+> orchestration truth. They are listed here so `/forge:update` removes the
+> orphaned files from `.forge/workflows/` (the `/forge:rebuild` regeneration
+> only clears manifest entries, it does not delete files on disk).
+
 For each found workflow file, check manifest status:
 ```sh
 node "$FORGE_ROOT/tools/generation-manifest.cjs" check .forge/workflows/{old-name}.md
@@ -1314,12 +1316,12 @@ Proceed to **Step 6**.
 node "$FORGE_ROOT/tools/banners.cjs" --phase 6 7 "Record state" drift
 ```
 
-> **Note:** `paths.forgeRoot` and `paths.forgeRef` were already written at the start
-> of Step 4. Step 6 does not repeat those writes — it records migration state only.
+> **Note:** `paths.forgeRef` was already written at the start
+> of Step 4. Step 6 does not repeat that write — it records migration state only.
 
 **Write `.forge/update-check-cache.json`** to record the completed migration.
 Read the existing file if present, update `migratedFrom`, `localVersion`,
-`distribution`, `forgeRoot`, `forgeRef`, `updateStatus`, `pendingReason`, and
+`distribution`, `forgeRef`, `updateStatus`, `pendingReason`, and
 `pendingMigrations`, then write it back. Use the Write or Edit tool — do not run
 a shell command for this step. The `.forge/` directory always exists at this
 point (it was checked earlier), so no `mkdir -p` is needed.
@@ -1330,7 +1332,6 @@ If the file does not exist, create it with:
   "migratedFrom": "<LOCAL_VERSION>",
   "localVersion": "<LOCAL_VERSION>",
   "distribution": "<DISTRIBUTION>",
-  "forgeRoot": "<FORGE_ROOT>",
   "forgeRef": "<LOCAL_VERSION>",
   "updateStatus": "complete",
   "pendingReason": null,

package/dist/forge-payload/hooks/preflight-session.cjs

@@ -0,0 +1,99 @@
+#!/usr/bin/env node
+'use strict';
+
+// Forge hook: preflight-session (FORGE-S27-T01 / item A1).
+//
+// SessionStart hook that primes the orchestrator preflight cache blob
+// (.forge/cache/preflight-status.json) by running forge-preflight.cjs once.
+//
+// SCOPING NOTE: SessionStart fires BEFORE any command is invoked and its
+// envelope carries no per-command signal, so this hook is deliberately
+// command-name-INDEPENDENT. Authoritative scoping to run-task / fix-bug /
+// run-sprint lives in the orchestration command-preamble path (which reads the
+// blob). This hook only ever:
+//   - is a STRICT NO-OP when .forge/ is absent (never changes behavior for
+//     non-Forge projects or unrelated commands);
+//   - primes the blob when .forge/ is present;
+//   - is FRESHNESS-GUARDED (idempotent): if an existing blob already matches
+//     the current config mtime + MASTER_INDEX hash, it is left untouched;
+//   - FAILS OPEN: any error -> stderr warning + exit 0. A hook failure must
+//     never block session start.
+
+process.on('uncaughtException', (err) => {
+  try { process.stderr.write(`forge preflight-session: internal error (fail-open): ${err.message}\n`); } catch (_) {}
+  process.exit(0);
+});
+
+const fs = require('fs');
+const path = require('path');
+const { spawnSync } = require('child_process');
+
+// Read + discard the envelope on fd 0 (fail-open on any parse error).
+try { fs.readFileSync(0, 'utf8'); } catch (_) { /* envelope optional */ }
+
+// Strict no-op when .forge/ is absent.
+const forgeDir = path.join(process.cwd(), '.forge');
+const configPath = path.join(forgeDir, 'config.json');
+if (!fs.existsSync(forgeDir) || !fs.existsSync(configPath)) {
+  process.exit(0);
+}
+
+const cacheDir = path.join(forgeDir, 'cache');
+const blobPath = path.join(cacheDir, 'preflight-status.json');
+
+// Freshness guard (idempotency): if the existing blob was computed from the
+// current config mtime, leave it untouched. forge-preflight records configMtime
+// in the blob; MASTER_INDEX changes flow through masterIndexHash, which the
+// blob also records, but the config mtime is the cheap primary key here.
+function configMtimeMs() {
+  try { return fs.statSync(configPath).mtimeMs; } catch (_) { return null; }
+}
+
+try {
+  if (fs.existsSync(blobPath)) {
+    const existing = JSON.parse(fs.readFileSync(blobPath, 'utf8'));
+    const curMtime = configMtimeMs();
+    if (existing && typeof existing.configMtime === 'number' &&
+        curMtime !== null && existing.configMtime === curMtime) {
+      // Blob is current — no rewrite, no duplicated side effect.
+      process.exit(0);
+    }
+  }
+} catch (_) { /* corrupt blob -> fall through and recompute */ }
+
+// Resolve forge-preflight.cjs. Prefer CLAUDE_PLUGIN_ROOT (set by the runtime),
+// fall back to config.paths.forgeRoot.
+let forgeRoot = process.env.CLAUDE_PLUGIN_ROOT || null;
+if (!forgeRoot) {
+  try {
+    const cfg = JSON.parse(fs.readFileSync(configPath, 'utf8'));
+    forgeRoot = cfg && cfg.paths && cfg.paths.forgeRoot;
+  } catch (_) { /* leave null */ }
+}
+if (!forgeRoot) process.exit(0); // cannot locate the tool — fail open
+
+const tool = path.join(forgeRoot, 'tools', 'forge-preflight.cjs');
+if (!fs.existsSync(tool)) process.exit(0);
+
+// Run preflight and capture its single JSON blob.
+const res = spawnSync('node', [tool, '--path', process.cwd()], {
+  cwd: process.cwd(),
+  encoding: 'utf8',
+  timeout: 8000,
+});
+if (res.status !== 0 || !res.stdout) process.exit(0); // fail open
+
+// Validate it parses before writing — never half-write a corrupt blob.
+let blob;
+try { blob = JSON.parse(res.stdout); } catch (_) { process.exit(0); }
+if (!blob || typeof blob !== 'object') process.exit(0);
+
+try {
+  fs.mkdirSync(cacheDir, { recursive: true });
+  // Atomic-ish write: tmp + rename so a crash never leaves a partial blob.
+  const tmpPath = blobPath + '.tmp';
+  fs.writeFileSync(tmpPath, JSON.stringify(blob, null, 2), 'utf8');
+  fs.renameSync(tmpPath, blobPath);
+} catch (_) { /* fail open */ }
+
+process.exit(0);

package/dist/forge-payload/init/phases/phase-3-materialize.md

@@ -63,11 +63,24 @@ node "$FORGE_ROOT/tools/substitute-placeholders.cjs" \
 ```
 
 Output directories (managed by the tool's `SUBDIR_OUTPUT_MAP`):
-- `base-pack/commands/`  → `.claude/commands/<prefix-lowercased>/`
-- `base-pack/personas/`  → `.forge/personas/`
-- `base-pack/skills/`    → `.forge/skills/`
-- `base-pack/workflows/` → `.forge/workflows/`
-- `base-pack/templates/` → `.forge/templates/`
+- `base-pack/commands/`     → `.claude/commands/<prefix-lowercased>/`
+- `base-pack/personas/`     → `.forge/personas/`
+- `base-pack/skills/`       → `.forge/skills/`
+- `base-pack/workflows/`    → `.forge/workflows/`
+- `base-pack/templates/`    → `.forge/templates/`
+- `base-pack/workflows-js/` → `.claude/workflows/` (JS orchestration workflows — FORGE-S28-T01)
+
+### Step 3a — Record generated JS workflows in the generation manifest
+
+After `substitute-placeholders.cjs` writes the JS workflows, record them in the generation
+manifest so `/forge:rebuild` and `/forge:health` can detect stale or missing copies:
+
+```sh
+node "$FORGE_ROOT/tools/generation-manifest.cjs" record .claude/workflows/wfl-run-task.js
+node "$FORGE_ROOT/tools/generation-manifest.cjs" record .claude/workflows/wfl-run-sprint.js
+```
+
+If either `record` call exits non-zero, the file was not written — re-run Step 3 and retry.
 
 If `project-context.json` is absent or missing required keys, halt Phase 3:
 ```

package/dist/forge-payload/integrity.json

@@ -1,6 +1,6 @@
 {
-  "version": "1.0.10",
-  "generated": "2026-05-31",
+  "version": "1.2.14",
+  "generated": "2026-06-03",
   "note": "Tamper-evident only. Authoritative source: /forge:update from remote.",
   "files": {
     "commands/add-pipeline.md": "529a2fc01be49815efa2cf1147528827ff1bfc12ce2ac8663b5a3a9781f8682e",
@@ -9,27 +9,28 @@
     "commands/check-agent.md": "f92121cb150d4cf601654e25c28d0c13389039518d05433f1040d727054b127e",
     "commands/config.md": "e67185f98cf0b045890398083991b9d48f32bf0d53bafda1eaca3ce0337ce49c",
     "commands/enhance.md": "d28f6414119e84973809d32900bb7245b51f565aed86112bc7a994bc6941b547",
-    "commands/health.md": "29c3e5f9808d996f53a432c4b3aff1c3346d7a40fff54518d8ea2ac2c6ba9d12",
+    "commands/health.md": "8f8160d67c09eed9be12445674fc2ab76cf31d548c6268c71f9777dca30c4e47",
     "commands/init.md": "cf7a70d55f718304820c310465071d3eea9d14a43131623f46b686499f630d4d",
-    "commands/rebuild.md": "e6d1e0bc29fd21c4c9109d8f40045b782207dfa6db2d98a9bf8c7254a80f09c6",
+    "commands/rebuild.md": "9de3a3c1e41169e18758c8598e297c2f395f9daefdd9772783b9b676982cf1c2",
     "commands/remove.md": "0ca5ec94d52959afaadd74910bd2c81c9872b71c9cc1ce8b89e5af69c06b304f",
     "commands/repair.md": "22e21614eee31ee630ead3f587b0b948b8c537e54ecef3b65e7d819c55c0ebd4",
     "commands/report-bug.md": "af8a54bf8887b35e5c880898dd45783f6c2e80d3dc031d6479a6be613ac43053",
     "commands/search.md": "befc4ba9f3146e7599308b35b95ec315e16895c0c18c9c2bab7549619d79bf1c",
     "commands/status.md": "2fa82a0b11d493918781aec6d95c39411f84fa923662ff3152b6de0cc830d755",
-    "commands/update.md": "a7eac4362a927eae708ee3844ef8d920f3166d1eafde36323d2dd92789fbdec9",
+    "commands/update.md": "c8b69b43567fa5fc328bb3dde31f386d0f577bbcc65dafd525400ae2fb16b132",
     "agents/store-query-validator.md": "f4c3573edcf6e28809515705362df611806a805c5269404fb17e31433cf3a81c",
     "agents/tomoshibi.md": "0c1032df80dfc25a4f482b1276965b5318754d95ccde9f24820846d564e49a27",
     "hooks/check-update.cjs": "9607cc0c51bedcbc123f5ea1d7a492916f29ee1f838846eec64776d5df70bbea",
     "hooks/forge-permissions.cjs": "8de18fe7e6c9ee1751926275a37f4bb68a5a53090d366d2aa44157c106e7857e",
     "hooks/post-init.cjs": "b05d65e7192d126733995989998a0b480604e0cc152887e6d7d0807fba3cf70d",
     "hooks/post-sprint.cjs": "28ddbd2c49d8b746c35d23a90b02362bc684d188159536a971b2d650bb85f3a9",
+    "hooks/preflight-session.cjs": "7b1ad8f12f2a1bb94650e2b632da0ed21862071bf63f4e6452c88007be630fae",
     "hooks/triage-error.cjs": "e47fcb81fd0f99b8ecc449b38765cdc6929bde2461b5f1727d26bc31f76bfcfa",
     "hooks/validate-write.cjs": "3ec7bbbb2cc19f22c78da5a9e7b5031066f4b669c76a10c9f8a92a1a48fcea0a",
     "schemas/transitions/bug.json": "27b17da42d1cebffbd4f61ab3dcd432a0017aa71997d548ed80d22c2fc3fad6a",
     "schemas/transitions/sprint.json": "2e0a629396e687b0ca88e1814ac3e35d84533a5d55b25237ab67feddce3c9deb",
     "schemas/transitions/task.json": "4c71849747baa0d585756e57c56325ff6219c078178374e60815dff91e25e3e7",
-    "schemas/enum-catalog.json": "fe6c4b9d0b326a22ccbe2d2e6b8ea81c4f81ea827b9b46507a214cbeffcd45d3",
+    "schemas/enum-catalog.json": "14fa6c470292fb433a4e33503669f567bd53e3851b25000c05ddca0fdfdd6e58",
     "tools/verify-integrity.cjs": "3ec3c970dd3d7c3001f8f373bcc40556803eadd2fc2afafb14f1c232cba4cc3f"
   }
 }

package/dist/forge-payload/meta/fragments/tool-discipline.md

@@ -35,7 +35,7 @@ project config outside `.forge/store/` are fine to `read`/`grep` directly.
   Never construct artifact paths manually — the tool resolves them from entity IDs.
 - Use `forge_verify_apply` after applying edits to confirm changes landed on disk.
   If `unchanged` is non-empty, re-apply those edits.
-- Never `bash node "$FORGE_ROOT/tools/store-cli.cjs" ...` — use the named MCP tool instead.
+- Never `bash node .forge/tools/store-cli.cjs ...` — use the named MCP tool instead.
   The tool is schema-validated and shorter.
 - Workflow text saying `forge_store write sprint '<json>'` means: call the MCP tool
   `forge_store` with that 2-positional shape. Not a shell command.

package/dist/forge-payload/meta/personas/meta-architect.md

@@ -64,7 +64,7 @@ When generating a project-specific Architect persona, incorporate:
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" north
+node .forge/tools/banners.cjs north
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🗻 **{Project} Architect** — I hold the shape of the whole. I give final sign-off.`

package/dist/forge-payload/meta/personas/meta-bug-fixer.md

@@ -67,7 +67,7 @@ When generating a project-specific Bug Fixer, incorporate:
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" rift
+node .forge/tools/banners.cjs rift
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🍂 **{Project} Bug Fixer** — I find what has decayed and restore it.`

package/dist/forge-payload/meta/personas/meta-collator.md

@@ -42,9 +42,9 @@ invokes the generated tool or falls back to manual collation.
 
 ## Preferred Method
 
-Read `paths.forgeRoot` from `.forge/config.json` → set as `FORGE_ROOT`. Then run:
+Run the vendored collate tool:
 ```bash
-node "$FORGE_ROOT/tools/collate.cjs"
+node .forge/tools/collate.cjs
 ```
 
 ## Fallback Method
@@ -56,17 +56,17 @@ the same outputs following the collation algorithm in
 ## Generation Instructions
 
 When generating a project-specific Collator, incorporate:
-- Emit the runtime-read pattern exactly as shown above — do NOT substitute
-  `paths.forgeRoot` as a literal string at generation time. The `$FORGE_ROOT`
-  variable must remain in the generated file so the path resolves from
-  `.forge/config.json` when the workflow runs, not when it is generated.
+- Emit the vendored project-relative invocation exactly as shown above —
+  `node .forge/tools/collate.cjs`. The tools closure is vendored into
+  `.forge/tools/` by `/forge:rebuild`, so the path resolves at runtime
+  without any plugin-root lookup.
 - The project's language for invoking the tool
 - The store path (.forge/store/)
 - The project prefix for ID formatting
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" drift
+node .forge/tools/banners.cjs drift
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🍃 **{Project} Collator** — I gather what exists and arrange it into views.`

package/dist/forge-payload/meta/personas/meta-engineer.md

@@ -64,7 +64,7 @@ When generating a project-specific Engineer persona, incorporate:
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" forge
+node .forge/tools/banners.cjs forge
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🌱 **{Project} Engineer** — I plan and build. I do not move forward until the code is clean.`

package/dist/forge-payload/meta/personas/meta-orchestrator.md

@@ -65,7 +65,7 @@ When generating a project-specific Orchestrator, incorporate:
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" tide
+node .forge/tools/banners.cjs tide
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🌊 **{Project} Orchestrator** — I move tasks through their lifecycle. I do not do the work; I watch that it flows.`

package/dist/forge-payload/meta/personas/meta-supervisor.md

@@ -86,7 +86,7 @@ When generating a project-specific Supervisor persona, incorporate:
 
 **Persona block format** — every generated workflow for this persona must open by running the identity banner using the Bash tool:
 ```bash
-FORGE_ROOT=$(node -e "console.log(require('./.forge/config.json').paths.forgeRoot)") && node "$FORGE_ROOT/tools/banners.cjs" oracle
+node .forge/tools/banners.cjs oracle
 ```
 Use `--badge` for compact inline contexts. The plain-text fallback for non-terminal output is:
 `🌿 **{Project} Supervisor** — I review before things move forward. I read the actual code, not the report.`

package/dist/forge-payload/meta/tool-specs/store-cli.spec.md

@@ -10,7 +10,7 @@ needed.
 
 ## Inputs
 
-- `.forge/config.json` — paths (specifically `paths.forgeRoot` for schema resolution)
+- `.forge/config.json` — project paths and prefix
 - `.forge/schemas/*.schema.json` — canonical JSON Schema files (primary source)
 - `forge/schemas/*.schema.json` — in-tree source schemas (fallback for dogfooding)
 - CLI arguments — command, entity type, JSON payload, flags

package/dist/forge-payload/meta/workflows/_fragments/event-emission-schema.md

@@ -59,7 +59,7 @@ After the subagent returns, the orchestrator constructs the event from:
 The orchestrator then emits via:
 
 ```
-node "$FORGE_ROOT/tools/store-cli.cjs" emit {sprintId} '{complete-event-json}'
+node .forge/tools/store-cli.cjs emit {sprintId} '{complete-event-json}'
 ```
 
 ## Why no example record here

package/dist/forge-payload/meta/workflows/_fragments/friction-emit.md

@@ -42,7 +42,7 @@ fields (`model`, `provider`, `eventId`, timestamps, token counts) are
 at drain time.
 
 ```sh
-node "$FORGE_ROOT/tools/friction-emit.cjs" \
+node .forge/tools/friction-emit.cjs \
   --workflow {workflow-key} \
   --persona  {persona-noun} \
   --issue    skill_unused \

package/dist/forge-payload/meta/workflows/_fragments/iron-laws.md

@@ -21,7 +21,7 @@ An Iron Laws section MUST contain exactly these three bullets in this order:
 
 - {WORKFLOW_SPECIFIC_LAW}
 - Read `.forge/personas/{persona}.md` first; print the persona identity line (emoji, name, tagline) to stdout before any other tool use.
-- All store I/O via `forge_store` (or `node "$FORGE_ROOT/tools/store-cli.cjs"`). Never edit `.forge/store/*.json` directly.
+- All store I/O via `forge_store` (or `node .forge/tools/store-cli.cjs`). Never edit `.forge/store/*.json` directly.
 ```
 
 ### Slot definitions

package/dist/forge-payload/meta/workflows/_fragments/progress-reporting.md

@@ -25,7 +25,7 @@ orchestrator monitors in real time.
 **Writing entries:** Use `store-cli progress`:
 
 ```
-node "$FORGE_ROOT/tools/store-cli.cjs" progress {sprintId} {agentName} {bannerKey} {status} "detail text"
+node .forge/tools/store-cli.cjs progress {sprintId} {agentName} {bannerKey} {status} "detail text"
 ```
 
 **Monitoring:** The orchestrator starts a Monitor on the progress log before
@@ -34,5 +34,5 @@ spawning each subagent and stops it after the subagent returns.
 **Clearing:** The orchestrator clears the progress log at task start:
 
 ```
-node "$FORGE_ROOT/tools/store-cli.cjs" progress-clear {sprintId}
+node .forge/tools/store-cli.cjs progress-clear {sprintId}
 ```

package/dist/forge-payload/meta/workflows/_fragments/store-cli-verbs.md

@@ -28,7 +28,7 @@ Notes for subagents:
   Do not `write` a task back with a new `status` field; the FSM is enforced
   by `update-status`. Syntax requires the field keyword `status` as the third
   argument — four args total:
-  `node "$FORGE_ROOT/tools/store-cli.cjs" update-status task {taskId} status {value}`
+  `node .forge/tools/store-cli.cjs update-status task {taskId} status {value}`
   The three-arg form `update-status task {taskId} {value}` is WRONG and will
   error. Always include `status` between the id and the value.
 - **`emit`** appends an event. There is no `append-event` / `add-event`.
@@ -38,6 +38,15 @@ Notes for subagents:
   canonical phase→filename map (so `set-summary {taskId} validation` just works).
   Never pass a hand-built `engineering/sprints/.../…-SUMMARY.json` path. Do not
   inline summaries into the entity via `write`.
+  - **`forge_store` tool shape (not CLI flags).** The tool has exactly two
+    fields: `command` (string) and `args` (positional string array). There is
+    NO `entity` / `id` / `phase` named field — passing them silently drops them.
+    The summary call is `forge_store({ command:"set-summary", args:["<id>", "<phase>"] })`
+    where `args[0]` is the record id and `args[1]` is the LITERAL phase key
+    (`plan`, `review_plan`, `implementation`, `code_review`, `validation`,
+    `triage`, `approve`). `args[1]` is NEVER the record id and NEVER a path —
+    putting the id in both slots is the canonical failure and the phase-ownership
+    guard rejects it with `expected summary key '<phase>'`.
 - **Artifact I/O:** Use `forge_artifact` for ALL phase artifact reads and writes
   (PLAN.md, PROGRESS.md, *-SUMMARY.json, CODE_REVIEW.md, etc.). Never construct
   artifact file paths manually — the tool resolves paths from entity IDs and
@@ -56,7 +65,7 @@ Notes for subagents:
   These spellings are parsed literally by tools (`preflight-gate.cjs`,
   `collate.cjs`) — do not invent new spellings or rename them in prose.
 - If you need a verb not on this list, run
-  `node "$FORGE_ROOT/tools/store-cli.cjs" --help` before improvising.
+  `node .forge/tools/store-cli.cjs --help` before improvising.
 - If you supply an unknown verb, entity type, enum value, or field name,
   store-cli appends a **Did you mean?** suggestion to the error message.
   Suggestions use Levenshtein distance (≤ 2) and a curated drift map for

package/dist/forge-payload/meta/workflows/meta-approve.md

@@ -31,7 +31,7 @@ The Architect gives final sign-off on a completed task after Supervisor approval
 
 - Approve only when the implementation is consistent with the project's architecture and the deployment posture is understood. Architectural sign-off is not a rubber stamp — it is the last point at which cross-cutting concerns can be caught cheaply.
 - Read `.forge/personas/architect.md` first; print the persona identity line (emoji, name, tagline) to stdout before any other tool use.
-- All store I/O via `forge_store` (or `node "$FORGE_ROOT/tools/store-cli.cjs"`). Never edit `.forge/store/*.json` directly.
+- All store I/O via `forge_store` (or `node .forge/tools/store-cli.cjs`). Never edit `.forge/store/*.json` directly.
 
 ## Store-Write Verification
 
@@ -42,9 +42,8 @@ The Architect gives final sign-off on a completed task after Supervisor approval
 ```
 
 0a. Pre-flight Gate Check:
-   - Resolve FORGE_ROOT (`node -e "console.log(require('./.forge/config.json').paths.forgeRoot)"`).
    - **Entity-mode resolution:** read the kickoff arguments. `--task {id}` → `entity_kind = "task"`, `record_id = {id}`. `--bug {id}` → `entity_kind = "bug"`, `record_id = {id}`. All store-cli calls below substitute `{entity_kind}` and `{record_id}` for the literal "task"/{taskId} placeholders.
-   - Run: `node "$FORGE_ROOT/tools/preflight-gate.cjs" --phase approve --{entity_kind} {record_id}`
+   - Run: `node .forge/tools/preflight-gate.cjs --phase approve --{entity_kind} {record_id}`
    - Exit 1 (gate failed) → print stderr and HALT. Do not proceed; do not attempt to produce the artifact.
    - Exit 2 (misconfiguration) → print stderr and HALT.
    - Exit 0 → continue.
@@ -53,7 +52,7 @@ The Architect gives final sign-off on a completed task after Supervisor approval
    - If `--force` is present in the invocation arguments, skip this step entirely.
    - If `entity_kind == "bug"`, skip this step entirely (bug state is managed by meta-fix-bug.md).
    - Read current task state:
-     `node "$FORGE_ROOT/tools/store-cli.cjs" read task {record_id} --json`
+     `node .forge/tools/store-cli.cjs read task {record_id} --json`
    - Extract the `status` field from the JSON output.
    - Allowed states for this phase: `review-approved`.
    - If the current status is NOT in the allowed set:
@@ -86,7 +85,7 @@ The Architect gives final sign-off on a completed task after Supervisor approval
 
 4. Finalize:
    - Transitions:
-     - **Task mode** — Update status: `node "$FORGE_ROOT/tools/store-cli.cjs" update-status task {taskId} status approved`. The status IS the verdict signal for task-mode commit gate (`STATUS_SOURCE` in `read-verdict.cjs`).
+     - **Task mode** — Update status: `node .forge/tools/store-cli.cjs update-status task {taskId} status approved`. The status IS the verdict signal for task-mode commit gate (`STATUS_SOURCE` in `read-verdict.cjs`).
      - **Bug mode** — NO status write. The bug remains `in-progress`. The verdict signal travels through `summaries.approve.verdict` written in step 5 below (read by `read-verdict.cjs § BUG_PHASE_VERDICT_SOURCE`). Writing `bug.status` here — especially writing `approved` or `verified` — violates `meta-fix-bug.md § Iron Laws #2` and is the trap that produced the FORGE-BUG-002 regression.
    - **Do NOT emit a phase event yourself.** The orchestrator (or kickoff handler) owns event emission — it composes the canonical event from runtime telemetry (model, provider, tokens, wall times) plus the SUMMARY you write in the next step. Subagents that call `store-cli emit` for phase events hallucinate runtime facts (see Plan 11 / Slice 2). Write the SUMMARY and return.
 
@@ -106,11 +105,11 @@ The Architect gives final sign-off on a completed task after Supervisor approval
    - Call (task mode) — optional for tasks, since `task.status` is the canonical signal.
      The sidecar path is auto-resolved from the record's `path` — never pass it:
      ```
-     node "$FORGE_ROOT/tools/store-cli.cjs" set-summary {taskId} approve
+     node .forge/tools/store-cli.cjs set-summary {taskId} approve
      ```
      Or (bug mode) — REQUIRED for bugs, this is the canonical verdict signal:
      ```
-     node "$FORGE_ROOT/tools/store-cli.cjs" set-bug-summary {bugId} approve
+     node .forge/tools/store-cli.cjs set-bug-summary {bugId} approve
      ```
    - In bug mode, if the set-bug-summary call exits non-zero, fix the sidecar JSON and retry. Do not return without a valid summary — the downstream commit gate has no other way to read the approval verdict.
 ```

package/dist/forge-payload/meta/workflows/meta-bug-triage.md

@@ -37,7 +37,7 @@ and the only phase that records the route decision read by the orchestrator
   reproduction has no business going to plan-fix or implement.
 - Read `.forge/personas/bug-fixer.md` first; print the persona identity
   line (emoji, name, tagline) to stdout before any other tool use.
-- All store I/O via `forge_store` (or `node "$FORGE_ROOT/tools/store-cli.cjs"`).
+- All store I/O via `forge_store` (or `node .forge/tools/store-cli.cjs`).
   Never edit `.forge/store/*.json` directly.
 - **Triage NEVER writes `bug.status`.** The orchestrator (`meta-fix-bug.md`)
   owns the `reported → triaged` and `triaged → in-progress` transitions.
@@ -60,8 +60,7 @@ and the only phase that records the route decision read by the orchestrator
 ```
 
 0. Pre-flight Gate Check:
-   - Resolve FORGE_ROOT (`node -e "console.log(require('./.forge/config.json').paths.forgeRoot)"`).
-   - Run: `node "$FORGE_ROOT/tools/preflight-gate.cjs" --phase triage --bug {bugId}`
+   - Run: `node .forge/tools/preflight-gate.cjs --phase triage --bug {bugId}`
    - Exit 1 (gate failed) → print stderr and HALT. Do not proceed.
    - Exit 2 (misconfiguration) → print stderr and HALT.
    - Exit 0 → continue.
@@ -70,7 +69,7 @@ and the only phase that records the route decision read by the orchestrator
    - Read `.forge/personas/bug-fixer.md` first; print the persona identity
      line to stdout before any other tool use.
    - Read the bug record:
-     `forge_store({ command:"read", entity:"bug", id:"{bugId}" })`
+     `forge_store({ command:"read", args:["bug", "{bugId}"] })`
    - Read business domain docs relevant to the reported symptom.
    - store-cli verbs: `read` | `list` | `write` | `emit` |
      `update-status` | `set-summary` | `set-bug-summary` | `describe` |
@@ -138,13 +137,17 @@ and the only phase that records the route decision read by the orchestrator
 
    - Call:
      ```
-     forge_store({ command:"set-bug-summary", entity:"bug",
-                   id:"{bugId}", phase:"triage" })
+     forge_store({ command:"set-bug-summary", args:["{bugId}", "triage"] })
      // sidecar path auto-resolved from the bug record's `path` — never pass it
      ```
-   - If the set-bug-summary call exits non-zero, fix the sidecar JSON
-     and retry (up to 3 attempts per the Store-Write Verification rule).
-     Do not proceed without a valid summary.
+     `forge_store` has only `command` + `args` (positional) — no
+     `entity`/`id`/`phase` field. `args[0]` is the bug id, `args[1]` is the
+     LITERAL phase key `triage` (never the bug id, never a path). See
+     `_fragments/store-cli-verbs.md`.
+   - If the set-bug-summary call exits non-zero (phase-ownership guard:
+     `expected summary key 'triage'`), `args[1]` was wrong — set it to `triage`
+     and retry (up to 3 attempts per the Store-Write Verification rule). Do not
+     proceed without a valid summary.
 
 > **Field-naming caution — runtime-tested.** The route field is named
 > `route`, never `path`. The bug schema's top-level `path` field is the