forgedock 1.0.22 → 1.1.2

package/commands/work-on/close.md

@@ -12,9 +12,11 @@ argument-hint: [issue number] [--repo GH_REPO] [--gh-flag GH_FLAG] [--pr PR_NUMB
 **Invoked by**: `work-on.md` Phase 6–7, after `review.md` returns `REVIEW_RESULT: status: COMPLETE`.
 **Output**: Update project board, close issue, update parent tracker, post trajectory log. Return final summary.
 
-**Agent model policy**: Default `model: "sonnet"`. If Sonnet is rate-limited, fall back to `model: "opus"`.
+**Agent model policy**: `model: "haiku"`, `effort: low` (mechanical tier — label transitions, annotation posting, board updates). Fallback: `model: "sonnet"` if rate-limited. Feature gate: pass `effort` only on Claude Code >= 2.1.154.
 **NEVER use plan mode (EnterPlanMode).**
 
+<!-- FORGE:SPEC_LOADED — work-on/close.md loaded and active. Agent is bound by this spec. -->
+
 ---
 
 ## Inputs
@@ -179,13 +181,15 @@ ISSUE_STATE=$(gh issue view {NUMBER} {GH_FLAG} --json state --jq '.state')
 
 **If `REMAINING_AFTER > 0`** (multi-phase: uncompleted phases remain):
 ```bash
-# Post phase-complete marker — router state 10 (MULTI_PHASE) will re-detect this issue
+# Post phase-complete marker — work-on.md's Universal continuation rule will re-read labels and continue to the next phase
 gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:PHASE:COMPLETE -->
 Phase complete. PR #{PR_NUMBER} merged to \`{PR_BASE}\`. ${REMAINING_AFTER} phase item(s) remain — leaving issue open for next pipeline iteration."
 
-# Update labels to reflect phase complete (not yet fully merged)
+# Update labels to reflect phase complete — remove current-phase labels and set next-phase label
+# so the router has a signal for which phase comes next (fixes: issue #1381)
 gh issue edit {NUMBER} {GH_FLAG} \
-  --remove-label "workflow:in-review,workflow:building,workflow:investigating" 2>/dev/null || true
+  --remove-label "workflow:in-review,workflow:building,workflow:investigating" \
+  --add-label "workflow:investigating" 2>/dev/null || true
 
 # EXIT — do not close, do not post trajectory, do not run C3–C6
 # Return CLOSE_RESULT: status: PHASE_COMPLETE to caller
@@ -444,6 +448,129 @@ Where:
 
 ---
 
+## Phase C5.2: Memory Index Update <!-- Added: forge#1316 -->
+
+**Goal**: Append this run's learnings to the per-repo memory index so future `investigate` runs can retrieve relevant priors. This is the write side of the compounding intelligence loop.
+
+**This phase is non-blocking** — if Gist creation or update fails, log the reason and continue to Phase C5.5. Never stall close for memory.
+
+**Skip if**: Terminal state is `INVALID` (nothing useful to learn from an invalid issue) OR `<!-- FORGE:TRAJECTORY -->` was already posted AND `<!-- FORGE:MEMORY_INDEXED -->` comment exists on the issue (idempotency guard).
+
+### Step 1: Compose memory entry
+
+Extract key fields from the pipeline:
+
+```bash
+ROOT_CAUSE=$(gh api repos/{GH_REPO}/issues/{NUMBER}/comments \
+  --jq '.[] | select(.body | contains("FORGE:INVESTIGATOR")) | .body' \
+  | sed -n '/### Root Cause/{n;p;q}' | head -1 | cut -c1-120)
+
+AFFECTED_FILES_BRIEF=$(gh api repos/{GH_REPO}/issues/{NUMBER}/comments \
+  --jq '.[] | select(.body | contains("FORGE:INVESTIGATOR")) | .body' \
+  | sed -n '/### Affected Files/{n;p;q}' | head -1 | cut -c1-120)
+
+DOMAIN_TAGS=$(gh issue view {NUMBER} {GH_FLAG} --json labels \
+  --jq '[.labels[].name | select(test("^(auth|billing|database|security|payments|gdpr|perf|ui|api|config)"))] | join(",")' 2>/dev/null || echo "")
+
+MEMORY_TIMESTAMP=$(date -u +"%Y-%m-%dT%H:%M:%SZ")
+
+MEMORY_ENTRY="MEMORY_ENTRY: issue={NUMBER} title=\"{TITLE}\" domain=\"${DOMAIN_TAGS}\" root_cause=\"${ROOT_CAUSE}\" outcome=\"${CARD_STATUS}\" files=\"${AFFECTED_FILES_BRIEF}\" lesson=\"${ROOT_CAUSE}\" timestamp=${MEMORY_TIMESTAMP}"
+```
+
+### Step 2: Find or create the memory index Gist
+
+```bash
+MEMORY_INDEX_ID=$(gh gist list --limit 100 \
+  --jq '.[] | select(.description | contains("FORGE:MEMORY_INDEX: {GH_REPO}")) | .id' 2>/dev/null | head -1)
+
+if [ -z "$MEMORY_INDEX_ID" ]; then
+  # First run — create the index Gist
+  TMPFILE=$(mktemp --suffix=.md)
+  cat > "$TMPFILE" <<GIST_EOF
+# ForgeDock Memory Index — {GH_REPO}
+<!-- FORGE:MEMORY_INDEX: {GH_REPO} -->
+Generated by ForgeDock close phase. Each line is a prior pipeline run.
+
+${MEMORY_ENTRY}
+GIST_EOF
+  # Memory gists MUST be secret — never pass --public here (forge#1587).
+  # The entry content below embeds real issue titles, root causes, and file
+  # paths; for a private consumer repo, --public would publish that content
+  # to a world-readable Gist. `gh gist create` is secret by default, so
+  # simply omitting --public is sufficient — the read side (investigate.md
+  # Phase 0.5, via `gh gist view`/`gh gist list`) works unchanged against
+  # secret gists for the authenticated owner.
+  MEMORY_INDEX_URL=$(gh gist create "$TMPFILE" \
+    --desc "FORGE:MEMORY_INDEX: {GH_REPO} — per-codebase learning index" \
+    2>/dev/null | head -1)
+  MEMORY_INDEX_ID=$(echo "$MEMORY_INDEX_URL" | sed 's|.*/||')
+  rm -f "$TMPFILE"
+  echo "[MEMORY] Created memory index Gist: ${MEMORY_INDEX_URL}"
+else
+  # Append to existing index
+  EXISTING_CONTENT=$(gh gist view "$MEMORY_INDEX_ID" 2>/dev/null)
+  UPDATED_CONTENT="${EXISTING_CONTENT}
+${MEMORY_ENTRY}"
+  INDEX_FILENAME=$(gh api gists/${MEMORY_INDEX_ID} --jq '.files | keys[0]' 2>/dev/null)
+  INDEX_FILENAME="${INDEX_FILENAME:-memory_index.md}"
+  TMPFILE=$(mktemp --suffix=.md)
+  echo "$UPDATED_CONTENT" > "$TMPFILE"
+  gh gist edit "$MEMORY_INDEX_ID" -f "$INDEX_FILENAME" "$TMPFILE" 2>/dev/null
+  EDIT_EXIT=$?
+  rm -f "$TMPFILE"
+  MEMORY_INDEX_URL="https://gist.github.com/${MEMORY_INDEX_ID}"
+  if [ $EDIT_EXIT -eq 0 ]; then
+    echo "[MEMORY] Appended to memory index: ${MEMORY_INDEX_URL}"
+  else
+    echo "WARNING: Failed to update memory index — run will not be persisted"
+    MEMORY_INDEX_URL=""
+  fi
+fi
+```
+
+### Migration: replacing an existing public memory-index Gist
+
+GitHub does not support flipping a Gist's visibility from public to secret in place. If you find an existing `FORGE:MEMORY_INDEX: {GH_REPO}` Gist that is **public** (check with `gh api gists/${MEMORY_INDEX_ID} --jq '.public'`), replace it:
+
+```bash
+# The gist ID here is the same one found by the FORGE:MEMORY_INDEX search above.
+OLD_PUBLIC_GIST_ID="$MEMORY_INDEX_ID"
+
+# 1. Save the existing content so no learnings are lost — and verify the save
+#    succeeded BEFORE deleting anything.
+gh gist view "$OLD_PUBLIC_GIST_ID" > /tmp/memory_index_migrate.md
+if [ ! -s /tmp/memory_index_migrate.md ]; then
+  echo "ABORT: failed to save gist content — not deleting the public gist." >&2
+  exit 1
+fi
+
+# 2. Delete the public gist (removes the world-readable copy)
+gh gist delete "$OLD_PUBLIC_GIST_ID" --yes
+
+# 3. Recreate it secret (no --public) with the same description tag so the
+#    next close/investigate run finds it via the FORGE:MEMORY_INDEX search
+gh gist create /tmp/memory_index_migrate.md \
+  --desc "FORGE:MEMORY_INDEX: {GH_REPO} — per-codebase learning index"
+```
+
+This is a one-time manual step per affected repo — Phase C5.2 itself only ever creates the index once (Step 2's `if [ -z "$MEMORY_INDEX_ID" ]` branch) and appends afterward, so a stale public gist is never auto-recreated secret on its own.
+
+### Step 3: Post audit annotation on the issue
+
+```bash
+if [ -n "$MEMORY_INDEX_URL" ]; then
+  gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:MEMORY_INDEXED -->
+This run has been indexed in the per-repo memory: ${MEMORY_INDEX_URL}
+
+Future \`investigate\` runs on \`{GH_REPO}\` will retrieve this entry as a prior when working on related issues.
+
+<!-- FORGE:MEMORY_INDEXED:COMPLETE -->"
+  echo "[MEMORY] Indexed issue #{NUMBER} into memory at: ${MEMORY_INDEX_URL}"
+fi
+```
+
+---
+
 ## Phase C5.5: Graph Decision Record (MANDATORY when PR exists)
 
 **Skip if**: `{PR_NUMBER}` is empty OR `<!-- FORGE:DECISION_RECORD -->` already posted on the PR.
@@ -577,7 +704,7 @@ CLOSE_RESULT:
   parent_closed: {true|false}
 ```
 
-Where `PHASE_COMPLETE` means the current phase was closed but uncompleted phases remain — the issue is left OPEN for the router to re-evaluate on the next pipeline iteration (state 10, MULTI_PHASE).
+Where `PHASE_COMPLETE` means the current phase was closed but uncompleted phases remain — the issue is left OPEN for work-on.md to re-evaluate on the next invocation (via Universal continuation rule: re-read labels, check terminal state, continue to next phase).
 
 ---
 

package/commands/work-on/decompose.md

@@ -12,7 +12,7 @@ argument-hint: [issue number] [--repo GH_REPO] [--gh-flag GH_FLAG]
 **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.
 
-**Agent model policy**: Default `model: "sonnet"`. If Sonnet is rate-limited, fall back to `model: "opus"`.
+**Agent model policy**: `model: "sonnet"` (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154.
 **NEVER use plan mode (EnterPlanMode).**
 
 ---
@@ -108,6 +108,24 @@ For each sub-issue, prepare:
 
 For each sub-issue (in dependency order):
 
+**Before creating each sub-issue, run the deterministic dedup check:** <!-- Added: forge#1335 -->
+
+```bash
+SUB_TITLE="{fix|feat|refactor}: {SUB_ISSUE_TITLE}"
+DEDUP_RESULT=$(scripts/issue-dedup.sh "$SUB_TITLE" {GH_FLAG} 2>&1)
+DEDUP_EXIT=$?
+if [ "$DEDUP_EXIT" -eq 1 ]; then
+  echo "DEDUP: Sub-issue near-duplicate detected — $DEDUP_RESULT"
+  echo "Skipping creation. Comment on the existing issue instead."
+  # Do NOT call gh issue create for this sub-issue
+elif [ "$DEDUP_EXIT" -eq 2 ]; then
+  echo "DEDUP: Usage error — $DEDUP_RESULT"
+  # Do NOT call gh issue create for this sub-issue — fix the invocation and retry
+fi
+```
+
+Only call `gh issue create` when the dedup script exits 0 (no match):
+
 ```bash
 gh issue create {GH_FLAG} \
   --title "{fix|feat|refactor}: {SUB_ISSUE_TITLE}" \

package/commands/work-on/investigate.md

@@ -11,9 +11,11 @@ argument-hint: [issue number] [--repo {owner}/{repo}] [--gh-flag "-R {owner}/{re
 
 Standalone investigation phase for the work-on pipeline. Validates whether an issue is real, determines root cause, posts a structured FORGE:INVESTIGATOR comment to GitHub, and updates workflow labels.
 
-**Agent model policy**: Default `model: "sonnet"`. If Sonnet is rate-limited, fall back to `model: "opus"`.
+**Agent model policy**: `model: "sonnet"` (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154.
 **NEVER use plan mode (EnterPlanMode).**
 
+<!-- FORGE:SPEC_LOADED — work-on/investigate.md loaded and active. Agent is bound by this spec. -->
+
 ---
 
 ## Inputs
@@ -27,6 +29,67 @@ If called from `work-on`, these are passed through. If invoked standalone, `--re
 
 ---
 
+## Phase 0.5: Memory Retrieval — Prior Run Priors <!-- Added: forge#1316 -->
+
+**Goal**: Before investigating, retrieve the top-k relevant prior pipeline runs from the per-repo memory index. Inject confirmed priors into the investigation context so the pipeline compounds intelligence across runs.
+
+**This phase is non-blocking** — if the memory index is absent or retrieval fails, log the reason and proceed to Phase 1A. Never stall the pipeline for memory.
+
+**Note on Gist visibility (forge#1587)**: The memory-index Gist is created **secret** (see `close.md` Phase C5.2 — no `--public` flag). `gh gist list` and `gh gist view` operate against the authenticated user's own Gists by description/id regardless of public/secret status, so the retrieval steps below work unchanged against a secret Gist. No code change is needed here — this note exists only so a future edit doesn't reintroduce a "must be public to be readable" assumption.
+
+### Step 1: Locate memory index Gist
+
+The memory index is a GitHub Gist tagged `<!-- FORGE:MEMORY_INDEX: {GH_REPO} -->`. Find it:
+
+```bash
+MEMORY_INDEX_URL=$(gh gist list --limit 100 \
+  --jq '.[] | select(.description | contains("FORGE:MEMORY_INDEX: {GH_REPO}")) | .url' 2>/dev/null | head -1)
+
+MEMORY_INDEX_ID=$(gh gist list --limit 100 \
+  --jq '.[] | select(.description | contains("FORGE:MEMORY_INDEX: {GH_REPO}")) | .id' 2>/dev/null | head -1)
+```
+
+If no memory index found: log `[MEMORY] No memory index for {GH_REPO} — starting fresh` and skip to Phase 1A.
+
+### Step 2: Retrieve relevant priors (top-k similarity)
+
+Read the memory index content:
+
+```bash
+MEMORY_CONTENT=$(gh gist view "$MEMORY_INDEX_ID" 2>/dev/null)
+```
+
+The memory index is a newline-delimited list of prior run entries, each formatted as:
+
+```
+MEMORY_ENTRY: issue={N} title="{TITLE}" domain="{DOMAIN_TAGS}" root_cause="{ROOT_CAUSE_SUMMARY}" outcome="{merged|invalid|blocked}" files="{AFFECTED_FILES}" lesson="{KEY_LESSON_ONE_LINE}" timestamp={ISO}
+```
+
+**Retrieve top-3 relevant priors** by matching against the current issue title and body:
+
+1. Extract keywords from the current issue title: lowercase, remove stop words, keep noun/verb tokens
+2. Score each memory entry: +2 for a file path overlap, +1 per keyword match in title/root_cause/lesson
+3. Return the top-3 highest-scoring entries (minimum score ≥ 1 to filter noise)
+
+If no entries score ≥ 1: log `[MEMORY] No relevant priors found` and skip to Phase 1A.
+
+### Step 3: Inject priors into investigation context
+
+For each retrieved prior, emit a structured block that Phase 1B can reference:
+
+```
+[MEMORY PRIOR #{RANK}]
+Issue: #{issue} — {title}
+Root cause: {root_cause}
+Outcome: {outcome}
+Key lesson: {lesson}
+Affected files: {files}
+```
+
+Print these blocks to stdout before Phase 1A begins. During Phase 1B (step 3 — blame analysis and step 5 — pickaxe pass), **explicitly check whether the current issue's suspected symbol or file appears in any prior root cause or affected files**. If a match is found, cite it in the FORGE:INVESTIGATOR comment's History Findings field as a `[MEMORY PRIOR]` hit.
+
+---
+
 ## Phase 1A: Load Issue & Check Resume State
 
 ```bash
@@ -67,10 +130,33 @@ The target repo is `{GH_REPO}` (resolved from `forge.yaml → project`). The wor
 
 If the issue specifies a **Code branch** (`**Code branch**: \`{branch}\``), check out that branch — the affected files may not be on the default branch.
 
+### Code Index Query (run BEFORE any grep exploration)
+
+If `scripts/code-index.sh` exists under `{REPO_PATH}`, query the pre-built symbol/import index first. This yields deterministic answers in one tool call and avoids redundant grep exploration across agents.
+
+```bash
+# Step 0A: Ensure index is current (cache-hit on unchanged HEAD — zero cost if already built)
+bash {REPO_PATH}/scripts/code-index.sh --repo-path {REPO_PATH} 2>/dev/null || true
+
+# Step 0B: Look up the symbol or file named in the issue (replace {SYMBOL} with the relevant name)
+bash {REPO_PATH}/scripts/code-index.sh query --symbol {SYMBOL} --repo-path {REPO_PATH} 2>/dev/null || true
+
+# Step 0C: Find all callers of that symbol
+bash {REPO_PATH}/scripts/code-index.sh query --callers {SYMBOL} --repo-path {REPO_PATH} 2>/dev/null || true
+
+# Step 0D: Find all importers of an affected file
+bash {REPO_PATH}/scripts/code-index.sh query --importers {AFFECTED_FILE} --repo-path {REPO_PATH} 2>/dev/null || true
+
+# Step 0E: Get all files in the affected domain (from issue labels/body)
+bash {REPO_PATH}/scripts/code-index.sh query --domain {DOMAIN_LABEL} --repo-path {REPO_PATH} 2>/dev/null || true
+```
+
+**Fallback**: If `scripts/code-index.sh` is absent or returns no results, proceed with standard grep exploration below. The index is an acceleration layer — its absence never blocks investigation.
+
 ### Investigation steps
 
 1. **Check the right branch** — read from the branch specified in the issue body (`**Code branch**: \`{branch}\``) if present
-2. **Read domain files** — start with the key files for the affected domain
+2. **Read domain files** — start with the key files for the affected domain (use index query results from Step 0E as the file list; fall back to directory inspection if index is absent)
 3. **Verify claims** — does the code actually have the problem described?
 4. **Git blame** — trace when/why the relevant code was written. Run bounded, local commands (no network round-trip):
    ```bash
@@ -106,6 +192,19 @@ If the issue specifies a **Code branch** (`**Code branch**: \`{branch}\``), chec
 
 The comment MUST include `<!-- INVESTIGATION:COMPLETE -->` at the very end, AFTER all required sections are present. This marker signals the investigation finished successfully.
 
+Before posting, resolve the attribution annotation link from `forge.yaml`:
+
+```bash
+ATTRIBUTION_ANNOTATION_LINK=$(grep -A5 "^attribution:" forge.yaml 2>/dev/null | grep "annotation_link:" | awk '{print $2}' | tr -d '"' || echo "false")
+ANNOTATION_LINK_FOOTER=""
+if [ "$ATTRIBUTION_ANNOTATION_LINK" = "true" ]; then
+  ANNOTATION_LINK_FOOTER="
+
+---
+*Pipeline powered by [ForgeDock](https://github.com/RapierCraftStudios/ForgeDock)*"
+fi
+```
+
 ```bash
 gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:INVESTIGATOR -->
 ## Investigation Report
@@ -146,6 +245,22 @@ gh issue comment {NUMBER} {GH_FLAG} --body "<!-- FORGE:INVESTIGATOR -->
 **{YES|NO}** — {reason}
 {if YES: proposed sub-issues with titles and dependencies}
 
+### Acceptance Spec
+{For each item in the issue's ## Acceptance Criteria section, emit one machine-checkable check line using the format below. If the issue has no Acceptance Criteria section, derive checks from the Recommendation above. Each check MUST be specific, observable, and testable — not vague prose. Checks are consumed by build/validate Phase B6.5 as the merge gate.}
+
+```
+ACCEPTANCE_CHECK: id={ac-1} type={exists|contains|command|behavior} target={file_path|command|url} matcher={string|exit_0|regex} description={one-line human description}
+ACCEPTANCE_CHECK: id={ac-2} type={exists|contains|command|behavior} target={file_path|command|url} matcher={string|exit_0|regex} description={one-line human description}
+```
+
+**Check types**:
+- `exists` — assert a file or directory exists (`target` = path, `matcher` = ignored)
+- `contains` — assert a file contains a string or regex (`target` = file path, `matcher` = string/regex)
+- `command` — run a shell command and assert exit 0 (`target` = shell command, `matcher` = `exit_0`)
+- `behavior` — assert a runtime/observable behavior via shell command (`target` = shell command, `matcher` = expected output string or regex)
+
+**Skipping**: if the issue has no verifiable acceptance criteria and none can be derived from the recommendation, emit a single sentinel: `ACCEPTANCE_CHECK: id=ac-skip type=skipped target=none matcher=none description=No machine-checkable criteria available — human review required`
+${ANNOTATION_LINK_FOOTER}
 <!-- INVESTIGATION:COMPLETE -->"
 ```
 

package/commands/work-on/review.md

@@ -12,9 +12,11 @@ argument-hint: [issue number] [--repo GH_REPO] [--gh-flag GH_FLAG] [--worktree P
 **Invoked by**: `work-on.md` Phase 4–5, after `build/validate.md` returns `GATE_PASSED: true`.
 **Output**: Push branch, create PR, invoke /review-pr --auto-merge, return result to caller.
 
-**Agent model policy**: Default `model: "sonnet"`. If Sonnet is rate-limited, fall back to `model: "opus"`.
+**Agent model policy**: `model: "sonnet"` (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154.
 **NEVER use plan mode (EnterPlanMode).**
 
+<!-- FORGE:SPEC_LOADED — work-on/review.md loaded and active. Agent is bound by this spec. -->
+
 ---
 
 ## Inputs
@@ -81,6 +83,30 @@ Do NOT push this branch. Human review required to identify the source of the mer
 fi
 ```
 
+## Phase R1: Non-Empty Commit Guard (MANDATORY — run before push) <!-- Added: forge#1305 -->
+
+Before pushing, verify the branch has at least one commit ahead of the PR base. This is the last-line defense against the phantom-commit hazard: a session that resumed from a partial FORGE:BUILDER comment (without `:COMPLETE`) would have skipped the commit step and could otherwise push an empty branch.
+
+```bash
+cd {WORKTREE_PATH}
+# Count commits on this branch that are not reachable from origin/{PR_BASE}
+COMMIT_COUNT=$(git rev-list --count HEAD ^origin/{PR_BASE} 2>/dev/null || echo "0")
+if [ "$COMMIT_COUNT" -eq 0 ]; then
+  gh issue comment {NUMBER} {GH_FLAG} --body "## Push Blocked — No Commits Ahead of Base
+
+Branch \`{BRANCH}\` has 0 commits ahead of \`origin/{PR_BASE}\`. Pushing this branch would create an empty PR.
+
+**Likely cause**: Build was interrupted after the FORGE:BUILDER comment was posted (implement.md Phase I6) but before the commit was created (validate.md Phase V5). The branch was pushed with no implementation on it.
+
+**Resolution**: Delete this branch, re-run \`/work-on {NUMBER}\` to restart the build phase. The partial FORGE:BUILDER comment (lacking \`FORGE:BUILDER:COMPLETE\`) will be detected and deleted, and the build will restart cleanly.
+
+<!-- FORGE:PUSH_BLOCKED_EMPTY_BRANCH -->"
+  gh issue edit {NUMBER} {GH_FLAG} --add-label "needs-human"
+  exit 1
+fi
+echo "Commit count ahead of origin/{PR_BASE}: $COMMIT_COUNT — OK to push"
+```
+
 ## Phase R1: Push Branch
 
 ```bash
@@ -122,7 +148,21 @@ Derive from issue title:
 - `docs(...):`  → `Docs: {description}`
 - fallback: use issue title as-is
 
-### R2B: Create PR
+### R2B: Resolve attribution footer (optional)
+
+Before creating the PR, check `forge.yaml → attribution.pr_footer`:
+
+```bash
+ATTRIBUTION_PR_FOOTER=$(grep -A5 "^attribution:" forge.yaml 2>/dev/null | grep "pr_footer:" | awk '{print $2}' | tr -d '"' || echo "false")
+```
+
+If `ATTRIBUTION_PR_FOOTER` is `true`, append the following footer to the PR body (once — never duplicate on retries):
+
+```
+> Orchestrated with [ForgeDock](https://github.com/RapierCraftStudios/ForgeDock) — state, scheduling, review, and memory on GitHub.
+```
+
+### R2C: Create PR
 
 ```bash
 gh pr create {GH_FLAG} \
@@ -146,18 +186,22 @@ gh pr create {GH_FLAG} \
 Closes #{NUMBER}
 
 **Implementation branch**: \`{BRANCH}\`
-**Base**: \`{PR_BASE}\`"
+**Base**: \`{PR_BASE}\`
+{IF_ATTRIBUTION_PR_FOOTER_TRUE:
+> Orchestrated with [ForgeDock](https://github.com/RapierCraftStudios/ForgeDock) — state, scheduling, review, and memory on GitHub.}"
 ```
 
 **Note**: `Closes #{NUMBER}` documents intent but does NOT auto-close for non-default-branch PRs. The close subcommand handles explicit closure after merge.
 
+**Attribution guard**: The footer line is appended once at PR creation. If the PR already exists (resume path), do NOT append the footer again — check the existing PR body first.
+
 If PR creation fails because a PR already exists for this branch:
 ```bash
 gh pr list {GH_FLAG} --head {BRANCH} --json number,url --jq '.[0]'
 ```
 Use the existing PR number and continue.
 
-### R2C: Update labels
+### R2D: Update labels
 
 ```bash
 gh issue edit {NUMBER} {GH_FLAG} \

package/commands/work-on-monolithic.md

@@ -1,6 +1,7 @@
 ---
 description: "[BENCHMARK] Monolithic work-on — single-prompt pipeline with no Skill sub-phase boundaries"
 argument-hint: [issue number]
+install: internal
 ---
 <!-- SPDX-FileCopyrightText: Copyright (c) RapierCraft Studios -->
 <!-- SPDX-License-Identifier: AGPL-3.0-or-later -->
@@ -11,9 +12,14 @@ argument-hint: [issue number]
 
 Orchestrator for the full issue lifecycle: investigate → build → review → merge → close. GitHub issues are the persistent context layer — read existing comments before starting, write structured reports back, use `workflow:*` labels to track state.
 
-**Agent model policy**: Default `model: "sonnet"`. Fallback: `model: "opus"` if Sonnet is rate-limited.
+**Agent model policy**: `model: "sonnet"` (standard tier). Fallback: `model: "opus"` if rate-limited. Feature gate: pass `effort` in Task/Skill spawns only on Claude Code >= 2.1.154.
 **NEVER use plan mode (EnterPlanMode).**
 
+> **Relationship to `/work-on` (canonical path)**: This file is a `[BENCHMARK]` reference variant. It documents the same inline-execution model that `/work-on` Phase 3 and `work-on/build.md` use as their **default** for STANDARD/fast-lane issues — all build sub-phases (context → architect → implement → validate) run inline in the current context window with no `Skill()` sub-agent spawns. The difference is that this file omits the worktree lifecycle, `FORGE:CONTRACT` posting, and the Spawn-Decision exception path for large builds. Use `/work-on` for production runs; use this file for baseline benchmarking or when comparing single-prompt token cost against the modular path. See `work-on/build.md` (Canonical Build Path section) for the reconciled build topology. <!-- Added: forge#1276 -->
+**NEVER use the Agent tool** — this spec executes all phases inline as a single monolithic prompt. The Agent tool would spawn a subprocess outside this context, breaking the single-prompt contract and losing all accumulated pipeline state.
+
+<!-- FORGE:SPEC_LOADED — work-on-monolithic.md loaded and active. Agent is bound by this spec. -->
+
 ### Compaction Resilience
 
 1. Write state to GitHub after EVERY significant step
@@ -64,7 +70,7 @@ gh api repos/{GH_REPO}/issues/{NUMBER}/comments --jq '.[] | {id: .id, author: .u
 
 **Check**: state (closed → STOP), terminal labels (`workflow:merged`/`workflow:invalid` → STOP), existing agent comments (`FORGE:INVESTIGATOR`, `FORGE:CONTRACT`, `FORGE:BUILDER`, `FORGE:TRAJECTORY`), parent tracker status.
 
-**Determine resume point**: No comments → Phase 1. Investigation exists + ready-to-build → Phase 3. Builder + no PR → Phase 4. Builder + PR open → Phase 5. PR merged + issue open → close issue.
+**Determine resume point**: No comments → Phase 1. Investigation exists + ready-to-build → Phase 3. Builder:COMPLETE + no PR → Phase 4. Builder without :COMPLETE (partial/interrupted build) + no PR → Phase 3 (partial-build cleanup). Builder + PR open → Phase 5. PR merged + issue open → close issue.
 
 **Classify lane**: Milestone → feature lane (`milestone/{slug}`). No milestone → fast lane (`staging`).