@hanzlaa/rcode 3.6.14 → 3.6.16

package/rihal/templates/UAT.md

@@ -1,29 +1,80 @@
-# UAT — {{phase_name}}
+# UAT.md — User Acceptance Test Template
 
-**Phase:** {{phase_id}}
-**Generated:** {{date}}
+Each phase produces a UAT.md alongside its SUMMARY.md. UAT documents how a non-developer can verify the phase delivered what it promised.
 
-## How to use
+## Frontmatter
 
-For each scenario below, mark status: ✅ pass | ❌ fail | ⏸ skipped |
-🔍 needs-human. Add notes inline.
+```yaml
+---
+phase: NN
+phase_name: <slug>
+generated: <ISO date>
+status: pending | passed | partial | diagnosed | blocked | skipped | human_needed
+last_run: <ISO date>
+---
+```
 
-## Scenarios
+## Sections
 
-### Scenario 1: {{golden path title}}
+### Phase goal (verbatim from ROADMAP.md)
 
-**Steps:**
-1. {{action}}
-2. {{action}}
+One paragraph copied straight from the phase's roadmap entry. No paraphrase.
 
-**Expected:** {{outcome}}
-**Status:** ⏸
-**Notes:**
+### What to verify
 
-### Scenario 2: {{edge case title}}
+A bullet list. Each bullet is one observable behavior:
 
-(Repeat as needed.)
+- ✅ **<feature>** — open <screen / endpoint / file>, do <action>, observe <expected outcome>
+- ✅ **<feature>** — run `<exact command>`, expect output containing `<exact substring>`
 
-## Outstanding items
+Every bullet must be runnable without reading the source code.
 
-Items left as ⏸ or 🔍 carry forward to the next milestone audit.
+### Setup steps
+
+If the verifier needs to seed data, set env vars, or start services:
+
+```bash
+# 1. install
+npx @hanzlaa/rcode install
+# 2. seed
+node .rihal/bin/rihal-tools.cjs ...
+# 3. ready
+open http://localhost:3000/foo
+```
+
+### Per-criterion result
+
+After running, fill in:
+
+| # | Criterion | Result | Notes |
+|---|-----------|:------:|-------|
+| 1 | <copy from "What to verify"> | ✅ / ⚠️ / ❌ | <one line> |
+| 2 | ... | ... | ... |
+
+### Outstanding items
+
+Anything that could not be verified, or where the verifier hit a blocker:
+
+- **<criterion>**: status = `blocked` because <reason>; needs <action> from <owner>.
+
+### Sign-off
+
+```
+Verified by: <name>
+Date: <ISO>
+Verdict: PASSED | PARTIAL | FAILED
+```
+
+## Status values
+
+- `pending` — UAT.md generated but not yet run.
+- `passed` — every criterion green.
+- `partial` — some criteria green, some not. Note which.
+- `diagnosed` — failed criteria have a known cause + a fix is in flight.
+- `blocked` — failed criteria need external work (data, infra, API access).
+- `skipped` — verifier intentionally did not run a criterion (with reason).
+- `human_needed` — automated check inconclusive; needs subjective human judgment.
+
+## Audit
+
+`/rihal:audit-uat` aggregates UAT.md files across all phases and surfaces outstanding items in one report.

package/rihal/templates/memory/INDEX.md

@@ -37,8 +37,8 @@
 - **`project/decisions.md`** — append every architectural choice as it's made
 - **`milestones/current.md`** — update on milestone phase transitions
 - **`incidents/known-issues.md`** — add when a workaround ships, remove when a real fix lands
-- **`distillates/`** — regenerate via `/rcode:memory-distill` when source files change
-- **Whole bank** — audit via `/rcode:memory-audit` at every milestone close
+- **`distillates/`** — regenerate via `/rihal-memory-distill` when source files change
+- **Whole bank** — audit via `/rihal-memory-audit` at every milestone close
 
 ## Constraints
 

package/rihal/templates/memory/distillates/project.distillate.md

@@ -1,11 +1,11 @@
 ---
 generated: true
 do-not-edit: true
-regenerate-with: /rcode:memory-distill
+regenerate-with: /rihal-memory-distill
 ---
 
 # Project Distillate — `{{PROJECT_NAME}}`
 
 > Generated, lossless compression of `project/`, `people/`, `milestones/current.md`, and `incidents/known-issues.md`. Optimised for LLM context loading. Hand-edits are overwritten on regenerate.
 
-_(Run `/rcode:memory-distill` to populate.)_
+_(Run `/rihal-memory-distill` to populate.)_

package/rihal/templates/memory/distillates/stack.distillate.md

@@ -1,11 +1,11 @@
 ---
 generated: true
 do-not-edit: true
-regenerate-with: /rcode:memory-distill
+regenerate-with: /rihal-memory-distill
 ---
 
 # Stack Distillate — `{{PROJECT_NAME}}`
 
 > Generated, lossless compression of `project/stack.md` only. Useful when you need stack context but not full project history.
 
-_(Run `/rcode:memory-distill --target stack` to populate.)_
+_(Run `/rihal-memory-distill --target stack` to populate.)_

package/rihal/templates/memory/project/stack.md

@@ -1,6 +1,6 @@
 # Stack — `{{PROJECT_NAME}}`
 
-Inventory of languages, frameworks, libraries, and services. Refreshed via `/rcode:memory-update` when stack changes.
+Inventory of languages, frameworks, libraries, and services. Refreshed via `/rihal-memory-update` when stack changes.
 
 ---
 

package/rihal/templates/summary.md

@@ -1,60 +1,77 @@
+# SUMMARY.md Template
+
+Each plan or sprint task produces a SUMMARY.md when it completes. The summary is the unit of work's lasting record — what was done, why, what remains, what other work it affects.
+
+## Frontmatter
+
+```yaml
 ---
-phase: "{{phase_id}}"
-status: complete
-closed_at: "{{closed_at}}"
+phase: NN
+plan: NN.M             # if applicable
+sprint: N              # if applicable
+story_id: NN.M         # if dev-story produced this
+generated: <ISO date>
+status: complete | partial | blocked
+verified: false        # flips to true after /rihal:verify-work passes
 ---
+```
+
+## Sections
+
+### One-liner
+
+A single line describing what was accomplished. Goes into `/rihal:progress` recent-work bullets, council recaps, milestone summaries. Be specific:
+
+> ❌ "Added some dashboard improvements"
+> ✅ "Wired live engagement metrics to dashboard top-3 tweets card; updates every 30s via SSE"
 
-# Phase Summary — {{phase_name}}
+### Outcomes
 
-<!-- P2: Omit any section that has nothing substantive to say. An empty section is worse than no section. -->
+What changed as a result of this work, observable from outside:
 
-**Phase ID:** {{phase_id}}
-**Closed:** {{closed_at}}
-**Sprint:** SPRINT.md
+- ...
 
-## Outcomes
+### Files touched
 
-- {{what was built, in user-visible terms}}
+```
+src/dashboard/Card.tsx        new
+src/dashboard/index.tsx       modified — added engagement subscription
+src/lib/sse.ts                new — SSE client wrapper
+test/dashboard.test.ts        new — 4 tests
+```
 
-## Decisions Made
+### Decisions
 
-<!-- Omit if no decisions were made that aren't already in SPRINT.md -->
-- {{key choices and their rationale}}
+Decisions made during execution that aren't already in `state.decisions[]`:
 
-## Deviations from Plan
+- **<decision>**: chose A over B because <one-line reason>. Tradeoff: <what we gave up>.
 
-<!-- Omit if execution matched the plan exactly -->
-- {{anything that diverged from SPRINT.md and why}}
+### Issues encountered
 
-## Issues Encountered
+Problems hit during execution and how they were resolved:
 
-<!-- Omit if no tasks were pruned, escalated, or skipped -->
-- {{tasks pruned, escalated, or skipped}}
+- **<problem>**: <one-line description>. Resolved by <action>. Time spent: <approx>.
 
-## Patterns Established
+If a problem was NOT resolved, it goes here AND becomes a follow-up filed below.
 
-<!-- Omit if no new patterns were introduced. List architectural/coding patterns future phases should follow. -->
-<!-- Example: "All service errors now wrap in ServiceError(code, message) — see auth.service.ts:42" -->
-- {{new pattern introduced by this phase and where it lives}}
+### Follow-ups
 
-## Provides
+GH issues filed (or to file) for work that this story surfaced but didn't ship:
 
-<!-- What this phase exposes for future phases to build on. Be specific: function names, API endpoints, data models, config keys. -->
-<!-- Example: "UserRepository.findByEmail() — src/repositories/user.repo.ts" -->
-- {{what future phases can reuse from this phase}}
+- #N — <title>
+- (to file) — <description>
 
-## Requires
+### Verification
 
-<!-- What this phase consumed from prior phases. Helps trace dependency chains. -->
-<!-- Example: "Database connection from Phase 3 (src/db/connection.ts)" -->
-- {{what this phase depended on from earlier phases}}
+How a verifier confirms this story is actually done:
 
-## Affects
+- Run `<command>`, expect `<output>`
+- Open `<UI element>`, click `<button>`, observe `<state change>`
 
-<!-- Downstream phases or components that may be impacted by what changed here. -->
-<!-- Example: "Phase 9 (checkout flow) — depends on Cart model introduced here" -->
-- {{phases or components that should re-verify after this phase}}
+This block feeds into UAT.md when the phase completes.
 
-## Hand-off
+### Status
 
-Next phase: {{next phase id and goal}}
+- **complete** — all acceptance criteria green, follow-ups filed.
+- **partial** — some AC green, others deferred. Document which AC are open.
+- **blocked** — couldn't proceed. Owner of unblock + ETA.

package/rihal/templates/verification-report.md

@@ -1,28 +1,83 @@
-# Verification Report — {{phase_name}}
+# Verification Report Template
 
-**Phase:** {{phase_id}}
-**Verified:** {{date}}
-**Verifier:** rihal-verifier
+Used by `/rihal:verify-phase`. Goal-backward analysis of whether the phase actually delivered the goal stated in ROADMAP.md.
 
-## Goal-backward check
+## Frontmatter
 
-**Original goal:** {{phase goal from PLAN.md}}
+```yaml
+---
+phase: NN
+phase_name: <slug>
+verifier: <agent / human>
+verified_at: <ISO date>
+verdict: pass | fail | partial
+goal_source: ROADMAP.md (commit-sha-at-phase-start)
+---
+```
 
-**Met?** ✅ / ◐ partial / ❌ not met
+## Sections
 
-**Evidence:**
-- {{file:line or test name proving the goal is met}}
+### Phase goal (verbatim)
 
-## Per-task verification
+> <paste the goal statement from ROADMAP.md exactly as written when this phase was planned>
 
-| # | Task | Status | Evidence |
-|---|------|--------|----------|
-| 1 | {{task}} | ✅ | {{evidence}} |
+### Goal decomposition
 
-## Gaps
+The goal as one sentence is too coarse to verify against. Decompose into checkable claims:
 
-- {{what wasn't delivered + why}}
+1. **Claim 1**: <one observable thing the goal implies>
+2. **Claim 2**: ...
+3. **Claim N**: ...
 
-## Recommendation
+### Evidence per claim
 
-- proceed / revise / block
+For each claim, list:
+
+- **Source of truth**: code path / file / endpoint / behavior
+- **Verification method**: command / test name / manual step
+- **Result**: ✅ / ❌ / ⚠️
+- **Notes**: anything surprising
+
+```
+| # | Claim | Method | Result | Notes |
+|---|-------|--------|:------:|-------|
+| 1 | ... | ... | ✅ | ... |
+```
+
+### Coverage gaps
+
+What the phase goal said but no evidence-trail covered. List explicitly:
+
+- **<claim>** has no implementation that can be pointed to. Either:
+  - The goal was aspirational and the phase didn't implement it (gap to file as follow-up)
+  - The implementation exists but no test or doc covers it (gap to file as test debt)
+
+### Side-effect audit
+
+What the phase changed BEYOND the goal. Often these are silent:
+
+- New dependencies added
+- Schema changes
+- API contract changes
+- File-tree changes (new dirs, deleted files)
+
+Each side effect should be intentional and documented.
+
+### Verdict reasoning
+
+One paragraph explaining why the verdict (pass / fail / partial) was chosen. Include any judgment calls.
+
+### Follow-ups
+
+Issues filed (or to file) for gaps surfaced during verification:
+
+- #N — <title>
+- (to file) — <description>
+
+## Sign-off
+
+```
+Verifier: <name or agent id>
+Date: <ISO>
+Verdict: <verdict from frontmatter>
+```

package/rihal/workflows/autonomous.md

@@ -18,7 +18,7 @@ interpos audit (issue #221) — DO NOT regress.
 2. **NEVER skip the methodology chain on greenfield projects.** Before
    the phase loop runs, the prerequisite check (next step) MUST verify:
    - `.planning/prd.md` exists (else halt → /rihal-create-prd)
-   - ROADMAP.md has milestone structure (else halt → /rihal-create-milestone)
+   - ROADMAP.md has milestone structure (else halt → /rihal-new-milestone)
    - `.planning/epics.md` exists (else halt → /rihal-create-epics-and-stories)
    See issue #219 + #229.
 
@@ -68,7 +68,7 @@ If `SKIP_FLAG=false` AND any prerequisite is missing, HALT with a clear message:
 
 The autonomous flow assumes a project that has already gone through:
   1. /rihal-create-prd               → produces .planning/prd.md
-  2. /rihal-create-milestone         → produces ROADMAP.md with M1..Mn
+  2. /rihal-new-milestone         → produces ROADMAP.md with M1..Mn
   3. /rihal-create-epics-and-stories → produces .planning/epics.md
   4. THEN /rihal-autonomous           ← you are here
 

package/rihal/workflows/create-prd.md

@@ -21,5 +21,5 @@ Skill not installed — run: npx @hanzlaa/rcode install
 ## On Completion
 
 /rihal-validate-prd — validate the PRD for completeness
-/rihal-create-milestone — build the milestone roadmap from the PRD
+/rihal-new-milestone — build the milestone roadmap from the PRD
 /rihal-edit-prd — revise the PRD

package/rihal/workflows/diagnose-issues.md

@@ -1,34 +1,87 @@
-# Workflow: rihal-diagnose-issues
-
 <purpose>
-Reusable diagnosis subroutine called from /rihal-verify-work when
-verification fails. Walks symptom → hypothesis → evidence → minimal
-repro and returns a structured finding. Not a top-level command.
+Diagnose issues surfaced during /rihal:verify-work. For each failed acceptance criterion, identify root cause, classify severity, and decide whether to fix in-flight or file as follow-up.
+
+Spawned by /rihal:verify-work when verification fails. Not typically called directly by users.
 </purpose>
 
-## Step 1 — Symptom capture
+<required_reading>
+@.rihal/references/output-format.md
+@.rihal/references/common-bug-patterns.md
+@.rihal/references/verification-patterns.md
+</required_reading>
+
+<process>
+
+## 1. Receive failure list
+
+`/rihal:verify-work` passes a structured list of failures:
+
+```json
+[
+  {
+    "criterion": "Auth login completes in <2s",
+    "expected": "redirect to /dashboard within 2s of submit",
+    "actual": "redirected after 8s",
+    "evidence": "browser network log shows 6s spent on /api/me"
+  }
+]
+```
+
+## 2. For each failure — diagnose
+
+Parallel-spawn one investigator per failure. Each investigator:
+
+a. **Reproduce** locally using the verification steps from UAT.md.
+b. **Bisect** — narrow down which file/function/commit introduced the regression. Use `git log --since` + `git bisect` if needed.
+c. **Classify** — bug / config / docs / environment / spec.
+d. **Severity** — critical (blocks ship) / high (degrades core feature) / medium (workaround exists) / low (cosmetic).
+e. **Root cause** — one-sentence explanation.
+f. **Fix proposal** — one of:
+   - **In-flight** — small enough to fix in this phase before merging
+   - **Follow-up** — substantial; file an issue and proceed
+   - **Stop** — phase shouldn't ship; halt and re-plan
+
+## 3. Aggregate diagnosis
+
+Each investigator returns a structured diagnosis. The orchestrator collates:
+
+```
+Failed criteria: 3
+Diagnosis:
+  Critical: 1   (must fix before ship)
+  High:     1   (recommended fix in-flight)
+  Medium:   1   (file as follow-up, ship anyway)
+```
+
+## 4. Decide
+
+- All criticals fixable in-flight → loop into rihal-code-fixer with the diagnosis bundle.
+- Any critical not fixable in-flight → return STOP to verify-work; phase doesn't ship.
+- Otherwise → file follow-ups, return PROCEED to verify-work.
 
-Record:
-- failing assertion / log line / wrong output
-- file:line where divergence first appears
-- last commit before symptom appeared (`git log -S<token>`)
+## 5. Output
 
-## Step 2 — Hypotheses
+`DIAGNOSIS.md` per failure, written to `.planning/phases/<NN>/diagnosis/<criterion-slug>.md`. Each contains:
 
-List 2–3 plausible causes. For each:
-- a one-line predicted mechanism
-- the cheapest test that would falsify it
+- Reproduction steps
+- Bisect result
+- Root cause
+- Severity + classification
+- Fix proposal + estimated effort
 
-## Step 3 — Evidence
+Plus an aggregate `DIAGNOSIS-INDEX.md` summarizing all failures and the orchestrator's decision.
 
-Run the falsifying tests in cheapest-first order. Stop at the first
-hypothesis that matches all evidence.
+</process>
 
-## Step 4 — Minimal repro
+## Success Criteria
 
-Reduce to the smallest input that still triggers the bug. Save under
-`.planning/diagnostics/<slug>.md`.
+- [ ] Every failed criterion has a DIAGNOSIS.md file
+- [ ] Each DIAGNOSIS.md has reproduction + root cause + fix proposal
+- [ ] Critical failures either get in-flight fix or halt the phase
+- [ ] Follow-up issues filed for medium/low items proceed without blocking ship
 
-## Step 5 — Hand-off
+## On Error
 
-Return JSON: `{ root_cause, repro_path, suggested_fix }` to the caller.
+- **Cannot reproduce** — mark diagnosis status `cannot-reproduce`. Often signals environment drift between verifier and dev. File as follow-up; do not block ship if all other criticals pass.
+- **Bisect inconclusive** — fall back to a hypothesis-driven investigation (read recent commits in the affected file, look for likely causes).
+- **Repair budget exhausted** — STOP. Return control to user with the full diagnosis bundle.

package/rihal/workflows/discuss-phase.md

@@ -152,7 +152,7 @@ Phase number from argument (required).
 ```bash
 INIT=$(node ".rihal/bin/rihal-tools.cjs" init phase-op "${PHASE}")
 if [[ "$INIT" == @file:* ]]; then INIT=$(cat "${INIT#@file:}"); fi
-AGENT_SKILLS_ADVISOR=$(node ".rihal/bin/rihal-tools.cjs" agent-skills rihal-advisor 2>/dev/null)
+AGENT_SKILLS_ADVISOR=$(node ".rihal/bin/rihal-tools.cjs" agent-skills rihal-advisor-researcher 2>/dev/null)
 ```
 
 Parse JSON for: `commit_docs`, `phase_found`, `phase_dir`, `phase_number`, `phase_name`, `phase_slug`, `padded_phase`, `has_research`, `has_context`, `has_plans`, `has_verification`, `plan_count`, `roadmap_exists`, `planning_exists`, `response_language`.

package/rihal/workflows/edit-prd.md

@@ -27,5 +27,5 @@ Skill not installed — run: npx @hanzlaa/rcode install
 ## ▶ Next Up
 
 - /rihal-validate-prd
-- /rihal-create-milestone
+- /rihal-new-milestone
 - /rihal-create-epics-and-stories

package/rihal/workflows/health.md

@@ -223,16 +223,16 @@ If `ORPHANS > 0`: `⚠ WARN — ${ORPHANS} orphaned worktree-agent-* artifact(s)
 
 **Action:** Count all pass/fail/warn results and display overall status.
 
-Total: `{N}/10 checks passed`
+Total: `{N}/9 checks passed`
 
-If all 10 pass:
+If all 9 pass:
 ```
-✓ All systems nominal — rihal is healthy
+✓ All systems nominal — rihal is healthy (9/9)
 ```
 
-If fewer than 10 pass:
+If fewer than 9 pass:
 ```
-⚠ {N}/10 checks passed — {M} issue(s) found
+⚠ {N}/9 checks passed — {M} issue(s) found
 Run: /rihal-update to repair installation issues
 Run: /rihal-status for project-state issues
 Run: /rihal-audit worktrees --prune to clean orphaned executor artifacts
@@ -240,7 +240,7 @@ Run: /rihal-audit worktrees --prune to clean orphaned executor artifacts
 
 ## Success Criteria
 
-- [ ] All 10 checks executed (skip state checks if no state.json)
+- [ ] All 9 checks executed (skip state checks if no state.json)
 - [ ] Each check result printed clearly
 - [ ] Final summary shows pass/fail count
 - [ ] Repair instructions shown if any checks fail

package/rihal/workflows/help.md

@@ -231,7 +231,6 @@ init → new-project → plan → execute → next → status → ship
 | `/rihal-dev-story <file>` | Execute an approved STORY by writing tests + code per AC. |
 | (internal) `check-implementation-readiness` | Guard called by `/rihal-plan` and `/rihal-execute` to verify PRD + architecture aligned before build. |
 | `/rihal-create-architecture` | Write an Architecture Decision Record (ADR). |
-| `/rihal-create-ux-design` | Realize a UX design that informs architecture and implementation. |
 | `/rihal-correct-course` | Course-correct mid-sprint when major change is discovered. |
 | `/rihal-sprint-planning` | Compute capacity, prioritize stories, create SPRINT.md. |
 | `/rihal-sprint-status` | Sprint progress — stories, points, velocity, burndown. |

package/rihal/workflows/memory-audit.md

@@ -33,7 +33,7 @@ fi
 
 ## Halt conditions
 
-- Memory Bank not initialised → instruct to run `/rcode:memory-init` first
+- Memory Bank not initialised → instruct to run `/rihal-memory-init` first
 
 ---
 
@@ -77,7 +77,7 @@ For `change-records/`, `incidents/post-mortems/`, `milestones/archive/`:
 For each `distillates/*.distillate.md`:
 - Read frontmatter `source-digest`
 - Recompute digest of current source files (per `rcode-memory-distill` rules)
-- If mismatch, emit a `warn` finding suggesting `/rcode:memory-distill`
+- If mismatch, emit a `warn` finding suggesting `/rihal-memory-distill`
 
 ### Step 8 — Render report
 

package/rihal/workflows/memory-distill.md

@@ -20,7 +20,7 @@ Regenerate Memory Bank distillates with lossless compression. Idempotent.
 
 ## Halt conditions
 
-- `.rihal/memory/` missing → instruct to run `/rcode:memory-init` first
+- `.rihal/memory/` missing → instruct to run `/rihal-memory-init` first
 - All sources empty (only template placeholders) → warn and exit; no point distilling empty content
 
 ---
@@ -70,7 +70,7 @@ For each target distillate, write:
 ---
 generated: true
 do-not-edit: true
-regenerate-with: /rcode:memory-distill
+regenerate-with: /rihal-memory-distill
 source-digest: <hash>
 generated-at: <ISO datetime>
 source-files:

package/rihal/workflows/memory-init.md

@@ -41,7 +41,7 @@ If `INDEX.md` exists:
 - List every file under `.rihal/memory/`
 - For each, count non-template lines (lines that don't start with `<!--` or contain `_(...)_` placeholders)
 - Print which files are empty / template-only and which are populated
-- Suggest `/rcode:memory-update` for surgical edits and exit
+- Suggest `/rihal-memory-update` for surgical edits and exit
 
 ### Step 2 — Copy templates
 
@@ -90,7 +90,7 @@ Add or update the `memory_bank` block in `.rihal/state.json`:
 Show:
 - File tree of `.rihal/memory/`
 - Files seeded vs files still empty
-- Suggested next command: `/rcode:memory-distill`
+- Suggested next command: `/rihal-memory-distill`
 
 ---
 

package/rihal/workflows/memory-update.md

@@ -15,7 +15,7 @@ Append-only surgical update to a single Memory Bank file. Never rewrites, never
 
 ## Preconditions
 
-- `.rihal/memory/` exists (run `/rcode:memory-init` first)
+- `.rihal/memory/` exists (run `/rihal-memory-init` first)
 
 ## Halt conditions
 
@@ -72,7 +72,7 @@ For `change-records/`:
 
 ### Step 5 — Suggest distillate refresh
 
-If the change was to `project/`, `milestones/current.md`, or `incidents/`, suggest `/rcode:memory-distill` to refresh distillates. Optional, not enforced.
+If the change was to `project/`, `milestones/current.md`, or `incidents/`, suggest `/rihal-memory-distill` to refresh distillates. Optional, not enforced.
 
 ---
 

package/rihal/workflows/research-phase.md

@@ -44,7 +44,7 @@ if [[ "$INIT" == @file:* ]]; then INIT=$(cat "${INIT#@file:}"); fi
 # If INIT is empty or INIT.ok is false: print "Error: rihal-tools init failed." and exit.
 # Extract: phase_dir, padded_phase, phase_number, state_path, requirements_path, context_path, response_language
 # If response_language is set, include "Respond in {value}." in all spawned subagent prompts.
-AGENT_SKILLS_RESEARCHER=$(node ".rihal/bin/rihal-tools.cjs" agent-skills rihal-researcher 2>/dev/null)
+AGENT_SKILLS_RESEARCHER=$(node ".rihal/bin/rihal-tools.cjs" agent-skills rihal-phase-researcher 2>/dev/null)
 ```
 
 ## Step 4: Spawn Researcher

package/rihal/workflows/scaffold-project.md

@@ -27,5 +27,5 @@ Skill not installed — run: npx @hanzlaa/rcode install
 ## ▶ Next Up
 
 - /rihal-create-prd
-- /rihal-create-milestone
+- /rihal-new-milestone
 - /rihal-do