forgedock 1.8.0 → 1.8.2

package/commands/work-on/decompose.md

@@ -1,20 +1,23 @@
 ---
 description: Decompose subcommand — break a complex issue into ordered sub-issues, post FORGE:DECOMPOSED, stop
-argument-hint: "[issue number] [--repo GH_REPO] [--gh-flag GH_FLAG]"
+context: fork
+argument-hint: "{NUMBER} --repo {GH_REPO} --gh-flag \"{GH_FLAG}\""
 ---
 <!-- SPDX-FileCopyrightText: Copyright (c) RapierCraft Studios -->
 <!-- SPDX-License-Identifier: AGPL-3.0-or-later -->
 
 # work-on/decompose — Decomposition Subcommand
 
+> **Skill Name Resolution (forked phase)**: `{FORGE_SKILL_PREFIX}` is the namespace this skill itself was invoked under — invoked as `forgedock:work-on:…` → `forgedock:` (nesting `:`); as `work-on:…` → empty (`install.sh`); as `forge-work-on-…` → `forge-` (Codex, nesting `-`); OpenCode → empty with `-` nesting. Confirm the target name in the available-skills list before calling it. A forked phase receives no resolved value from its caller; never guess, and if the target skill is not listed return BLOCKED "skill not found: <name>".
+
 **Input**: $ARGUMENTS
 
-**Invoked by**: `work-on.md` routing loop, when `INVESTIGATE_RESULT.decompose = YES`.
-**Output**: Create sub-issues, update parent tracker, post `<!-- FORGE:DECOMPOSED -->` comment, set labels. STOP — each sub-issue runs its own /work-on.
+**Invoked by**: the work-on router, when `INVESTIGATE_RESULT.decompose = YES`. This skill runs in an isolated forked context: it sees only its args and re-reads everything else from GitHub.
+**Output**: Create sub-issues, update parent tracker, post `<!-- FORGE:DECOMPOSED -->` comment, set `workflow:decomposed`. STOP — each sub-issue runs its own /work-on. This skill is the SOLE owner of the `workflow:decomposed` label. On BLOCKED it adds `needs-human` and posts the blocker.
 
-**Engine coverage** (forge#2379): this subcommand's `command` name (`work-on/decompose`) and completion marker (`FORGE:DECOMPOSED:COMPLETE`) are now registered as a real phase in the headless engine's phase table — `decompose` in `packages/protocol/src/phases.js`'s `PHASE_IDS`/`PHASE_MARKERS`, and the matching `decompose` entry in `bin/engine/phases.mjs`'s `PHASES` array. The engine's `investigate` phase hands off to it on `DECOMPOSE:YES` instead of terminating in place, so a headless `runIssue()` walk now actually dispatches this subcommand rather than stopping short.
+**Engine coverage** (forge#2379): this subcommand's `command` name (`work-on/decompose`) and completion marker (`FORGE:DECOMPOSED:COMPLETE`) are registered as a real phase in the headless engine's phase table — `decompose` in `packages/protocol/src/phases.js`'s `PHASE_IDS`/`PHASE_MARKERS`, and the matching `decompose` entry in `bin/engine/phases.mjs`'s `PHASES` array. The engine's `investigate` phase hands off to it on `DECOMPOSE:YES` instead of terminating in place, so a headless `runIssue()` walk dispatches this subcommand rather than stopping short.
 
-**Agent model policy**: `model: "{DEFAULT_MODEL}"` — resolved from forge.yaml `agents.default_model`, else "sonnet" (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154. This file's mechanical bits (label transitions, sub-issue creation) stay at this tier because they're interleaved with the reasoning-heavy sub-issue design steps in the same `Skill()` invocation — see `work-on.md` section "Model and Effort Tiering — What Actually Applies". <!-- Added: forge#1827 -->
+**Agent model policy**: `model: "{DEFAULT_MODEL}"` — resolved from forge.yaml `agents.default_model`, else "sonnet" (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154. This file's mechanical bits (label transitions, sub-issue creation) stay at this tier because they're interleaved with the reasoning-heavy sub-issue design steps in the same `Skill()` invocation. <!-- Added: forge#1827 -->
 **NEVER use plan mode (EnterPlanMode).**
 
 ---
@@ -23,8 +26,111 @@ argument-hint: "[issue number] [--repo GH_REPO] [--gh-flag GH_FLAG]"
 
 Parse from $ARGUMENTS:
 - `{NUMBER}` — issue number (required)
-- `--repo {GH_REPO}` — GitHub repo (e.g. `{owner}/{repo}` — resolved from `forge.yaml → project`)
-- `--gh-flag {GH_FLAG}` — gh CLI repo flag (e.g. `-R {owner}/{repo}`)
+- `--repo {GH_REPO}` — GitHub repo, e.g. `{owner}/{repo}` (required)
+- `--gh-flag {GH_FLAG}` — gh CLI repo flag, e.g. `-R {owner}/{repo}` (required)
+
+If `{NUMBER}`, `--repo` or `--gh-flag` is missing, print the result block with `status: BLOCKED`, blocker: "missing required arg: <name>", and STOP (no GitHub write is possible without them).
+
+---
+
+## Script resolution
+
+**Shell state does not persist between Bash tool calls** (each call is a fresh shell): paste the block below at the top of every bash command in this skill that calls `resolve_script` or uses `FORGE_ROOT`/`UNIVERSAL_DIR`.
+
+```bash
+# Canonical script resolution (keep byte-identical across specs; guarded by scripts/forge-root.test.sh).
+# Shell state does NOT persist between Bash tool calls: include this whole block at the top of
+# every command that uses resolve_script or FORGE_ROOT.
+REPO_PATH="${REPO_PATH:-$(yq '.paths.root // ""' forge.yaml 2>/dev/null)}"
+[ -n "$REPO_PATH" ] && [ "$REPO_PATH" != "null" ] || REPO_PATH="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
+ADAPTIVE_DIR_RAW="${REPO_PATH}/$(yq '.adaptive_scripts.directory // ".forgedock/scripts"' forge.yaml 2>/dev/null || echo '.forgedock/scripts')"
+ADAPTIVE_DIR=$(realpath -m "$ADAPTIVE_DIR_RAW" 2>/dev/null || echo "$ADAPTIVE_DIR_RAW")
+ADAPTIVE_ENABLED=$(yq '.adaptive_scripts.enabled // "true"' forge.yaml 2>/dev/null || echo 'true')
+# Bounds check: reject adaptive_scripts.directory values that escape the repo root.
+# Normalize REPO_PATH the same way ADAPTIVE_DIR is normalized (realpath -m) so a trailing
+# slash in paths.root does not inject a '//' into the glob and trigger a false positive.
+REPO_PATH_NORM=$(realpath -m "$REPO_PATH" 2>/dev/null || echo "$REPO_PATH")
+if [[ "$ADAPTIVE_DIR" != "${REPO_PATH_NORM}/"* ]]; then
+  echo "WARNING: adaptive_scripts.directory resolves outside repo root ('$ADAPTIVE_DIR') — adaptive tier disabled" >&2
+  ADAPTIVE_ENABLED=false
+fi
+# ForgeDock's own install root (holds scripts/ and commands/), resolved ONCE by the canonical bootstrap
+# below — never the consumer repo: plugin installs set neither FORGE_HOME nor FORGEDOCK_HOME, and a
+# repo-relative fallback would execute (or miss) a same-named script controlled by the consumer repo.
+# Resolution: $FORGEDOCK_HOME (authoritative when set) > $FORGE_HOME > $CLAUDE_PLUGIN_ROOT > the
+# ~/.claude/commands symlink target > the Claude Code plugin cache/marketplace dirs. <!-- forge#3098 -->
+# FORGE_ROOT bootstrap (canonical; keep byte-identical across specs, guarded by scripts/forge-root.test.sh)
+FORGE_ROOT=""
+if [ -n "${FORGEDOCK_HOME:-}" ]; then case "$FORGEDOCK_HOME" in /*) FORGE_ROOT="$FORGEDOCK_HOME" ;; esac; else
+  # Portable to bash 3.2 (macOS), BSD/GNU coreutils and zsh: no mapfile, no sort -V, no bare globs (zsh aborts on no match).
+  _l="$HOME/.claude/commands/work-on.md"; _l="$(readlink -f "$_l" 2>/dev/null || readlink "$_l" 2>/dev/null)"; [ -n "$_l" ] && _l="$(dirname "$(dirname "$_l")")"
+  # newest cached version first: numeric major.minor.patch of the version dir name only; a release outranks its pre-release (1.10.0 > 1.9.0 > 1.9.0-rc1)
+  _v="$(find -L "$HOME/.claude/plugins/cache" -mindepth 3 -maxdepth 3 -type d 2>/dev/null | awk -F/ '$(NF-1)=="forgedock"{v=$NF;p=index(v,"-");r=1;if(p){v=substr(v,1,p-1);r=0};split(v,a,".");printf "%d %d %d %d %s\n",a[1],a[2],a[3],r,$0}' | sort -k1,1nr -k2,2nr -k3,3nr -k4,4nr | cut -d' ' -f5-)"
+  _m="$(find -L "$HOME/.claude/plugins/marketplaces" -mindepth 1 -maxdepth 1 -type d -iname '*forgedock*' 2>/dev/null)"
+  # '${CLAUDE_PLUGIN_ROOT}' is substituted by Claude Code when it loads a plugin spec (the exact spelling only, never as an env var), so a running plugin resolves to its own root first; unsubstituted (other runtimes) it stays a literal that the /* check rejects.
+  _k="$(printf '%s\n' '${CLAUDE_PLUGIN_ROOT}' "${FORGE_HOME:-}" "$_l" "$_v" "$_m")"
+  while IFS= read -r _c; do
+    case "$_c" in /*) [ -z "$FORGE_ROOT" ] && [ -f "$_c/scripts/verify-phase-trail.sh" ] && [ -f "$_c/scripts/lint-dispatch-prompt.sh" ] && [ -f "$_c/scripts/is-docs-only.sh" ] && [ -f "$_c/bin/engine/resolve.mjs" ] && [ -f "$_c/bin/engine/orchestrate-canary.mjs" ] && [ -f "$_c/bin/engine/admission.mjs" ] && FORGE_ROOT="$_c" ;; esac
+  done <<< "$_k"
+fi
+UNIVERSAL_DIR="${FORGE_ROOT:+$FORGE_ROOT/scripts}"   # empty => tier 3 skipped, prose tier
+# NOTE: never resolve this via `which` or `find` — universal scripts are
+# install-relative, not installed on $PATH, so a PATH lookup always misses.
+# FORGE_ROOT (above) is the deterministic resolution; there is NO repo-path fallback.
+# Pipeline agents MUST NOT use `find` (unbounded or filesystem-wide) to
+# locate pipeline scripts under any circumstances: if UNIVERSAL_DIR/${operation}.sh
+# does not exist, resolve_script() falls through to Tier 4 (prose) below,
+# which is always safe and available. A missing script is never a reason
+# to search the filesystem. <!-- Added: forge#1984 -->
+
+resolve_script() {
+  local operation="$1"
+  # Tier 2: per-repo adaptive (skip if disabled)
+  if [ "$ADAPTIVE_ENABLED" != "false" ] && [ -f "${ADAPTIVE_DIR}/${operation}.sh" ]; then
+    echo "adaptive:${ADAPTIVE_DIR}/${operation}.sh"
+    return
+  fi
+  # Tier 3: universal script
+  if [ -n "$UNIVERSAL_DIR" ] && [ -f "${UNIVERSAL_DIR}/${operation}.sh" ]; then
+    echo "universal:${UNIVERSAL_DIR}/${operation}.sh"
+    return
+  fi
+  # Tier 4: prose fallback
+  echo "prose:"
+}
+```
+
+---
+
+## Result emission
+
+Every exit path of this skill — COMPLETE, ALREADY_DONE, every guard, every failure — ends by printing exactly one block as the final reply:
+
+```
+DECOMPOSE_RESULT:
+  status: COMPLETE | ALREADY_DONE | BLOCKED
+  sub_issues: [N, N, ...]
+  comment_url: {url of the FORGE:DECOMPOSED comment; empty if none}
+  blocker: {description if status=BLOCKED, else empty}
+```
+
+`sub_issues` is a plain list of issue numbers on every path (`[]` when none), never objects.
+
+### Blocked exit
+
+Every `BLOCKED` exit first runs this procedure (set `BLOCKER` to the reason text), then prints the result block with `status: BLOCKED`. It guarantees the parent never keeps a terminal `workflow:decomposed` label without sub-issues:
+
+```bash
+BLOCKER="{reason}"
+run() { if [ -n "${DRY_RUN:-}" ]; then echo "DRY_RUN: $*"; else "$@"; fi; }
+run gh issue edit {NUMBER} {GH_FLAG} --add-label "needs-human" --remove-label "workflow:decomposed" 2>/dev/null || true
+run gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:DECOMPOSE_BLOCKED -->
+## Decomposition Blocked
+
+${BLOCKER}
+
+Human attention required (\`needs-human\`)."
+```
 
 ---
 
@@ -40,9 +146,21 @@ gh api repos/{GH_REPO}/issues/{NUMBER}/comments \
   --jq '.[] | select(.body | contains("FORGE:INVESTIGATOR")) | .body'
 ```
 
-**Validation**:
-- If FORGE:INVESTIGATOR comment is absent → EXIT with `DECOMPOSE_RESULT: status: BLOCKED`, blocker: "No investigation report found — run investigate first"
-- If investigation report has no Decomposition Assessment section, OR the assessment does not list any sub-issues → EXIT with `DECOMPOSE_RESULT: status: BLOCKED`, blocker: "Investigation report has no decomposition plan — re-run investigate with explicit decomposition scope"
+**MANDATORY — Owner override detection**: After reading the investigation comment, read ALL other comments on the issue to check for owner override signals:
+
+```bash
+gh api repos/{GH_REPO}/issues/{NUMBER}/comments \
+  --jq '.[] | select(.body | contains("FORGE:INVESTIGATOR") | not) | {author: .user.login, body: .body}'
+```
+
+Scan non-agent comments for override signals — phrases like "do not", "do NOT", "instead", "revert", "remove this", "override", "actually", or explicit disagreement with the investigation's recommendation. If an override comment is found from a repo owner or admin (not a bot):
+
+1. **Document the override**: Note which direction the owner is steering (e.g., "remove the feature" vs. investigation's "keep with warnings")
+2. **Re-derive sub-issue scopes**: Derive sub-issue titles, bodies, and file scope from the override direction — NOT from the original investigation recommendation. The investigation's Decomposition Assessment may list sub-issues that are now stale or contradictory with the override.
+3. **If override makes a sub-issue obsolete**: Skip creating it. Note the skip reason in the decomposition comment (Phase D5).
+4. **If override changes the sequencing dependency**: Revise the execution order so that the override's primary action (e.g., "strip the feature") completes before any downstream doc/SDK sub-issues are built against it.
+
+**Why this matters**: Sub-issues scoped before an override are built against a stale premise. A docs sub-issue scoped as "neutralize liability language" becomes incorrect if the upstream schema sub-issue will fully remove the feature — the docs sub-issue should instead be "remove all references to the deleted feature." Building both in parallel against the pre-override scope produces contradictory staging state.
 
 > **Shared Scoping Convention**: This investigation gate (blocks without a `FORGE:INVESTIGATOR` comment) is the **reference pattern** for investigation-gated issue creation across the pipeline. `milestone.md` Step 4 and `orchestrate.md` MUST apply the same principle: read code and identify all affected call sites BEFORE writing any issue body. Sub-issue bodies created in Phase D3 MUST use the Pipeline Issue Template defined in `issue.md` Phase 3D — that template is the single canonical standard for all automated issue creation. <!-- Added: forge#293 -->
 
@@ -58,14 +176,40 @@ MILESTONE_TITLE=$(gh issue view {NUMBER} {GH_FLAG} --json milestone --jq '.miles
 
 ---
 
-## Phase D1: Resume Check
+## Phase D1: Guards and Resume Check
+
+Evaluate in this order. Each BLOCKED exit runs the Blocked exit procedure.
+
+**Guard 1 — issue is itself a sub-issue**: an issue whose body carries a `**Parent**: #N` reference is a child created by an earlier decomposition and is never decomposed again (prevents recursive decomposition). <!-- Added: forge#2379 -->
+
+```bash
+ISSUE_BODY=$(gh issue view {NUMBER} {GH_FLAG} --json body --jq '.body')
+PARENT_REF=$(printf '%s\n' "$ISSUE_BODY" | grep -E '^\*\*Parent\*\*: #[0-9]+' | head -1)
+[ -n "$PARENT_REF" ] && echo "SUB_ISSUE_GUARD: ${PARENT_REF}"
+```
+
+If `PARENT_REF` is non-empty → BLOCKED, blocker: "Issue #{NUMBER} is a sub-issue (${PARENT_REF}) — sub-issues are never decomposed again; run /work-on build on it directly".
+
+**Guard 2 — investigation report present**: If the FORGE:INVESTIGATOR comment is absent → BLOCKED, blocker: "No investigation report found — run investigate first".
+
+**Guard 3 — decomposition plan present**: If the investigation report has no Decomposition Assessment section, OR the assessment does not list any sub-issues → BLOCKED, blocker: "Investigation report has no decomposition plan — re-run investigate with explicit decomposition scope".
+
+**Resume check**:
 
 ```bash
 gh api repos/{GH_REPO}/issues/{NUMBER}/comments \
-  --jq '.[] | select(.body | contains("FORGE:DECOMPOSED")) | .body'
+  --jq '[.[] | select(.body | contains("<!-- FORGE:DECOMPOSED -->"))] | last // empty | {url: .html_url, body: .body}'
+```
+
+If a `<!-- FORGE:DECOMPOSED -->` comment exists → decomposition already complete. Derive `sub_issues` as a plain number list from that comment's `- #N:` lines and the comment URL from `.url`:
+
+```bash
+# $DECOMP_BODY is the .body of the comment found above
+SUB_LIST=$(printf '%s\n' "$DECOMP_BODY" | sed -nE 's/^- #([0-9]+).*/\1/p' | paste -sd, - | sed 's/,/, /g')
+echo "sub_issues: [${SUB_LIST}]"
 ```
 
-- If `<!-- FORGE:DECOMPOSED -->` comment exists → decomposition already complete. EXIT with `DECOMPOSE_RESULT: status: ALREADY_DONE`.
+Run Phase D6 (idempotent label re-assert), then EXIT with `DECOMPOSE_RESULT: status: ALREADY_DONE`, `sub_issues: [${SUB_LIST}]`, `comment_url` from the comment.
 
 ---
 
@@ -91,24 +235,26 @@ If `GIST_URLS` is non-empty, a `## Prior Investigation` section will be appended
 
 ## Phase D2: Design Sub-Issues
 
-From the Decomposition Assessment in the investigation report, extract:
+From the Decomposition Assessment in the investigation report (adjusted for any owner override detected in Phase D0), extract:
 1. Sub-issue titles (in dependency order — independent issues first)
 2. Dependencies between sub-issues (if issue B depends on issue A, A is created first)
 3. Brief description for each sub-issue body
 
 For each sub-issue, prepare:
-- **Title**: from investigation report's proposed sub-issue title, including the final `{fix|feat|refactor}: ` prefix used in Phase D3
+- **Title**: from investigation report's proposed sub-issue title (re-derived from the owner override when one was detected), including the final `{fix|feat|refactor}: ` prefix used in Phase D3
 - **Body**: brief description of scope + `**Parent**: #{NUMBER}` + dependency note if applicable
 - **Labels**: inherit priority label (P0/P1/P2) from parent; do NOT copy workflow labels
 - **Milestone**: same milestone title as parent (if parent has one)
 
+Sub-issues made obsolete by an owner override are not created; record each as `SKIPPED_NOTES` (title + reason) for the Phase D5 comment.
+
 **Ordering rule**: Create independent sub-issues first. If sub-issue B depends on A, create A first so its issue number can be referenced in B's body.
 
 ---
 
 ## Phase D2.5: Equivalent Set Check
 
-Before creating any sub-issue, check whether the complete planned title set already exists for this parent. This covers an interrupted earlier decomposition: its child issues may exist even though it never reached the `FORGE:DECOMPOSED` comment checked in D1.
+Before creating any sub-issue, check whether the complete planned title set already exists for this parent. This covers an interrupted earlier decomposition: its child issues may exist even though it never reached the `FORGE:DECOMPOSED` comment checked in Phase D1.
 
 ```bash
 # SUB_TITLES contains the final, ordered titles prepared in D2.
@@ -124,14 +270,19 @@ CANONICAL_SUB_ISSUES=$(jq --arg parent "**Parent**: #{NUMBER}" --argjson expecte
   | ($canonical | map(.title) | sort) as $actual
   | if $actual == $expected then $canonical else empty end
 ' <<< "$PARENT_SUB_ISSUES")
+
+# Plain number list for the result block, e.g. "101, 102"
+CANONICAL_NUMBERS=$(printf '%s' "$CANONICAL_SUB_ISSUES" | jq -r 'map(.number) | sort | map(tostring) | join(", ")' 2>/dev/null)
 ```
 
-If `CANONICAL_SUB_ISSUES` is non-empty, the decomposition has already created an equivalent set. Do **not** create issues, edit the parent body, or post another decomposition comment. Preserve those original sub-issues and EXIT with:
+If `CANONICAL_SUB_ISSUES` is non-empty, the decomposition has already created an equivalent set. Do **not** create issues, edit the parent body, or post another decomposition comment. Preserve those original sub-issues, run Phase D6 (idempotent label re-assert), and EXIT with:
 
 ```
 DECOMPOSE_RESULT:
   status: ALREADY_DONE
-  sub_issues: {CANONICAL_SUB_ISSUES}
+  sub_issues: [${CANONICAL_NUMBERS}]
+  comment_url:
+  blocker:
 ```
 
 If no exact title-set match exists, continue. A partial set is not equivalent: let `/issue` dedup protect its existing children while creating only the missing work.
@@ -140,12 +291,19 @@ If no exact title-set match exists, continue. A partial set is not equivalent: l
 
 ## Phase D3: Create Sub-Issues
 
-For each sub-issue (in dependency order), route creation through the `/issue` create-hook's programmatic invocation contract (`commands/issue.md` Programmatic Invocation Contract, added in #2085) instead of calling `gh issue create` directly. `/issue`'s Phase 2D runs the same `scripts/issue-dedup.sh` check this file used to run manually — the standalone pre-check below is removed; dedup is now enforced inside the create-hook on every call, with no bypass path. <!-- Changed: forge#2086 — route through /issue create-hook -->
+For each sub-issue (in dependency order), route creation through the `/issue` create-hook's programmatic invocation contract (`commands/issue.md` Programmatic Invocation Contract, added in #2085) instead of calling `gh issue create` directly. `/issue`'s Phase 2D runs the same `scripts/issue-dedup.sh` check this file used to run manually — dedup is enforced inside the create-hook on every call, with no bypass path. <!-- Changed: forge#2086 — route through /issue create-hook -->
+
+Initialise once before the loop: `CREATED_NUMBERS=""` (comma-separated, for the result block) and `SKIPPED_NOTES=""`.
 
-**Compose the sub-issue body to a temp file** (avoids quoting issues when passed as `--body-file`):
+**Sanitise the title and compose the sub-issue body to a temp file** (avoids quoting issues when passed as `--body-file`):
 
 ```bash
 SUB_TITLE="{fix|feat|refactor}: {SUB_ISSUE_TITLE}"
+# Defense-in-depth: /issue's arg tokenizer (commands/issue.md, forge#2094) uses
+# an xargs-based tokenizer that never expands backtick/$(...) substitution, so
+# this is no longer required for safety — but strip it anyway so the raw title
+# stays readable if it round-trips through any other eval-based consumer.
+SUB_TITLE=$(printf '%s' "$SUB_TITLE" | tr '`' "'" | sed 's/\$(/$ (/g')
 SUB_BODY_FILE="$(mktemp)"
 cat > "$SUB_BODY_FILE" <<'SUB_BODY_EOF'
 ## Problem
@@ -175,23 +333,52 @@ Files that need changes:
 SUB_BODY_EOF
 ```
 
+**Build the milestone arg** (only when the parent has a milestone — a parent with no milestone must not receive `--milestone ""`):
+
+```bash
+MILESTONE_ARG=""
+[ -n "$MILESTONE_TITLE" ] && MILESTONE_ARG=" --milestone \"${MILESTONE_TITLE}\""
+```
+
 **Invoke `/issue` in programmatic mode**:
 
 ```
-ISSUE_SKILL_OUTPUT=$(Skill(skill="issue", args="--title \"${SUB_TITLE}\" --body-file \"${SUB_BODY_FILE}\" --label \"{PRIORITY_LABEL}\" --milestone \"{MILESTONE_TITLE}\""))
+ISSUE_SKILL_OUTPUT=$(Skill(skill="{FORGE_SKILL_PREFIX}issue", args="--title \"${SUB_TITLE}\" --body-file \"${SUB_BODY_FILE}\" --label \"{PRIORITY_LABEL}\"${MILESTONE_ARG}"))
 ```
 
+If the skill is not found → BLOCKED, blocker: "skill not found: issue".
+
 `/issue` runs Phase 2D dedup, Phase 3F body validation, then creates the issue (Phase 4) — no separate pre-check needed on this side.
 
-**Extract the created sub-issue number from the Skill output** (see `commands/issue.md` Phase 4C/4E — it echoes `Created: {url}` and reports `**#{NUMBER}**: {title}`):
+If the `--milestone` flag fails (milestone not found by name): re-invoke without `${MILESTONE_ARG}`, set `MILESTONE_SKIPPED=1`, and note in the FORGE:DECOMPOSED comment that milestone assignment was skipped.
+
+**Extract the created sub-issue number from the Skill output** (see `commands/issue.md` Phase 4C/4E — it echoes `Created: {url}` and reports `**#{NUMBER}**: {title}`). If parsing yields nothing, fall back to an exact-title search scoped to this parent (covers `/issue` output not being captured by the Skill harness):
 
 ```bash
 # Match either the "Created: {url}" line (extract trailing /issues/N) or the "**#{NUMBER}**" bold report line.
 SUB_NUMBER=$(echo "$ISSUE_SKILL_OUTPUT" | grep -oE 'issues/[0-9]+' | head -1 | grep -oE '[0-9]+')
 [ -z "$SUB_NUMBER" ] && SUB_NUMBER=$(echo "$ISSUE_SKILL_OUTPUT" | grep -oE '\*\*#[0-9]+\*\*' | head -1 | grep -oE '[0-9]+')
 
+# Fallback: exact-title search, restricted to issues whose body carries this parent's reference
+# (so a pre-existing near-duplicate that /issue dedup stopped on is never mistaken for the new child).
+# Retries absorb GitHub Search API indexing lag.
+if [ -z "$SUB_NUMBER" ]; then
+  for _resolve_attempt in 1 2 3; do
+    SUB_NUMBER=$(gh issue list {GH_FLAG} --search "in:title \"${SUB_TITLE}\"" --state open --limit 10 --json number,title,body 2>/dev/null \
+      | jq -r --arg t "$SUB_TITLE" --arg p "**Parent**: #{NUMBER}" '[.[] | select(.title == $t and (.body | contains($p)))][0].number // empty')
+    [ -n "$SUB_NUMBER" ] && break
+    sleep 2
+  done
+fi
+
+rm -f "$SUB_BODY_FILE"
+
 if [ -z "$SUB_NUMBER" ]; then
   echo "WARNING: /issue did not report a created issue number for sub-issue '${SUB_TITLE}' — likely a Phase 2D dedup STOP (near-duplicate found) or a usage error. Skipping this sub-issue: do not reference it in the parent tracker or in dependent sub-issue bodies. Review the Skill output above and, if a near-duplicate exists, comment on the existing issue instead."
+  SKIPPED_NOTES="${SKIPPED_NOTES}
+- ${SUB_TITLE} — not created (dedup stop or /issue error)"
+else
+  CREATED_NUMBERS="${CREATED_NUMBERS:+${CREATED_NUMBERS}, }${SUB_NUMBER}"
 fi
 ```
 
@@ -221,9 +408,7 @@ Investigation findings from the parent issue are available as Knowledge Gists:
 fi
 ```
 
-Capture the created issue number from the output URL for the tracker checklist.
-
-If `--milestone` flag fails (milestone not found by name): omit the flag and note in the FORGE:DECOMPOSED comment that milestone assignment was skipped.
+After the loop, if `CREATED_NUMBERS` is empty (no sub-issue was created) → BLOCKED, blocker: "No sub-issues were created (all skipped or deduplicated) — review the /issue output".
 
 ---
 
@@ -245,7 +430,7 @@ TRACKER="
 - [ ] #{SUB_ISSUE_2_NUMBER} — {SUB_ISSUE_2_TITLE} _(depends on #{SUB_ISSUE_1_NUMBER})_
 ..."
 
-gh issue edit {NUMBER} {GH_FLAG} --body "${CURRENT_BODY}${TRACKER}"
+gh issue edit {NUMBER} {GH_FLAG} --body "${CURRENT_BODY}${TRACKER}" # allowlist:check-command-side-effects
 ```
 
 ---
@@ -259,8 +444,11 @@ SHOW_ATTRIBUTION=$(yq '.branding.show_attribution // "true"' forge.yaml 2>/dev/n
 > Pipeline powered by [ForgeDock](https://github.com/RapierCraftStudios/ForgeDock)"
 ```
 
+Compose the comment body. Include the `### Skipped Sub-Issues` section only when `SKIPPED_NOTES` is non-empty (each owner-override-obsolete sub-issue and each not-created sub-issue, with its reason), an `### Owner Override` line when Phase D0 detected one (the direction the owner steered), and a milestone-skipped note when `MILESTONE_SKIPPED=1`.
+
 ```bash
-gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:DECOMPOSED -->
+[ "${DRY_RUN:-false}" = "true" ] && { echo "[DRY_RUN] would post the comment below"; exit 0; }
+COMMENT_URL=$(gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:DECOMPOSED -->
 ## Decomposition Complete
 
 This issue has been broken into sub-issues. Each sub-issue runs through its own /work-on pipeline independently.
@@ -270,21 +458,39 @@ This issue has been broken into sub-issues. Each sub-issue runs through its own
 {for each sub-issue, in dependency order:}
 - #{SUB_ISSUE_NUMBER}: {TITLE}{if has dependency: _(depends on #{DEP_NUMBER})_}
 
+{if SKIPPED_NOTES non-empty:}
+### Skipped Sub-Issues
+
+${SKIPPED_NOTES}
+
+{if owner override detected:}
+### Owner Override
+
+{override direction and how it changed the sub-issue scopes/order}
+
 ### Decomposition Rationale
 
 {brief summary of why this issue was decomposed and the dependency ordering chosen}
 ${ATTRIBUTION_LINE}
-<!-- FORGE:DECOMPOSED:COMPLETE -->"
+<!-- FORGE:DECOMPOSED:COMPLETE -->") # allowlist:check-command-side-effects
 ```
 
 ---
 
 ## Phase D6: Update Labels
 
+Decompose is the SOLE owner of the `workflow:decomposed` label. Transition with tiered dispatch; a transition to `decomposed` removes every other workflow label. Also run this block (idempotently) on both ALREADY_DONE exits.
+
 ```bash
-gh issue edit {NUMBER} {GH_FLAG} \
-  --add-label "workflow:decomposed" \
-  --remove-label "workflow:ready-to-build,workflow:building,workflow:investigating" 2>/dev/null || true
+run() { if [ -n "${DRY_RUN:-}" ]; then echo "DRY_RUN: $*"; else "$@"; fi; }
+RESOLUTION=$(resolve_script 'transition-label'); TIER="${RESOLUTION%%:*}"; SCRIPT_PATH="${RESOLUTION#*:}"
+case "$TIER" in
+  adaptive|universal) run bash "$SCRIPT_PATH" {NUMBER} {GH_FLAG} decomposed ;;
+  prose)
+    run gh issue edit {NUMBER} {GH_FLAG} --add-label "workflow:decomposed" \
+      --remove-label "workflow:investigating,workflow:ready-to-build,workflow:building,workflow:in-review,workflow:awaiting-merge,workflow:merged,workflow:invalid" 2>/dev/null || true
+    ;;
+esac
 ```
 
 ---
@@ -293,15 +499,7 @@ gh issue edit {NUMBER} {GH_FLAG} \
 
 Decomposition is a terminal route for the parent issue. Each sub-issue will be picked up separately by /work-on.
 
-Return structured output to the router:
-
-```
-DECOMPOSE_RESULT:
-  status: COMPLETE | ALREADY_DONE | BLOCKED
-  sub_issues: [{number}, ...]
-  comment_url: {url of FORGE:DECOMPOSED comment}
-  blocker: {description if status=BLOCKED}
-```
+Print the result block (see Result emission) as the final reply, with `status: COMPLETE`, `sub_issues: [${CREATED_NUMBERS}]` (numbers only) and `comment_url: ${COMMENT_URL}`.
 
 **Router behavior after DECOMPOSE_RESULT**: `break` — do not continue to build/review/close for the parent issue.
 
@@ -314,7 +512,7 @@ The subcommand writes its results to GitHub (FORGE:DECOMPOSED comment + sub-issu
 ```
 DECOMPOSE_RESULT:
   status: COMPLETE | ALREADY_DONE | BLOCKED
-  sub_issues: [{number}, ...]
+  sub_issues: [N, N, ...]
   comment_url: {url of posted FORGE:DECOMPOSED comment}
   blocker: {description if status=BLOCKED}
 ```