@jenga-ai/agent 3.6.0 → 4.1.0

package/scripts/verify-postinstall-reconcile.sh

@@ -330,13 +330,27 @@ run_install "$PKG2C" "$CONSUMER_C" "$FIXTURE/case-v2.log"
 
 # Detect whether this filesystem is even case-insensitive; on a case-SENSITIVE fs
 # both names legitimately coexist and the old one is genuinely stale.
-if [ -f "$CONSUMER_C/.agents/skills/alpha/skill.md" ] && \
-   [ -f "$CONSUMER_C/.agents/skills/alpha/SKILL.md" ] && \
-   [ "$(cat "$CONSUMER_C/.agents/skills/alpha/skill.md" 2>/dev/null)" != "$(cat "$CONSUMER_C/.agents/skills/alpha/SKILL.md" 2>/dev/null)" ]; then
-  CASE_INSENSITIVE=0
-else
+#
+# This probes the filesystem DIRECTLY rather than inferring the answer from the
+# mirrored tree. The previous inference was circular -- it asked whether both
+# skill.md and SKILL.md exist with DIFFERENT content, but both fixture packages
+# above write the identical body ("# alpha"), so the inequality was false by
+# construction and the branch concluded "case-insensitive" on every filesystem.
+# On macOS that happened to be the right answer; on a case-SENSITIVE fs (any
+# Linux CI runner) it was wrong, and the case-insensitive branch below then ran
+# an assert_grep for "written-this-run" that can only hold when the two names
+# collide -- recording a FAIL that no individual @test asserts, so only
+# "harness reports zero failures overall" caught it. Surfaced by the E28_S17
+# mirror staging gate's first Linux runs (2026-09-28).
+CASE_PROBE="$FIXTURE/case-probe"
+mkdir -p "$CASE_PROBE"
+: > "$CASE_PROBE/probe"
+if [ -e "$CASE_PROBE/PROBE" ]; then
   CASE_INSENSITIVE=1
+else
+  CASE_INSENSITIVE=0
 fi
+rm -rf "$CASE_PROBE"
 
 if [ "$CASE_INSENSITIVE" -eq 1 ]; then
   # The whole point: the skill must still exist under SOME name after the upgrade.
@@ -353,8 +367,20 @@ if [ "$CASE_INSENSITIVE" -eq 1 ]; then
   assert_grep "written-this-run" "$FIXTURE/case-v2.log" \
     "identity guard reports the spared entry as written-this-run"
 else
-  pass "case-only rename (skipped: filesystem is case-sensitive)"
-  pass "case-only rename (skipped: filesystem is case-sensitive) (.claude)"
+  # These names must remain PREFIX-COMPATIBLE with the case-insensitive branch
+  # above: assert_check in tests/postinstall-delete-reconciliation.bats matches
+  # with grep -F on "<tab><name>", a substring match, so a trailing "(n/a ...)"
+  # qualifier still satisfies an assertion written against the canonical name
+  # while keeping the skip visible in the harness output. The identity-guard
+  # line below already followed this convention; the two case-only-rename lines
+  # did not -- they were renamed wholesale to "case-only rename (skipped: ...)",
+  # which shares no prefix with the canonical name, so on a case-sensitive fs
+  # the "case-only rename does not delete the file the run just wrote" @test
+  # would fail with "no such check recorded by the harness". That never fired
+  # before only because the broken detection above forced every filesystem down
+  # the case-insensitive branch.
+  pass "case-only rename does not delete the just-written file (.agents) (n/a: filesystem is case-sensitive, both names legitimately coexist)"
+  pass "case-only rename does not delete the just-written file (.claude) (n/a: filesystem is case-sensitive, both names legitimately coexist)"
   pass "identity guard reports the spared entry as written-this-run (n/a on case-sensitive fs)"
 fi
 

package/skills/index/scripts/board_index.py

@@ -6,6 +6,7 @@ import argparse
 import hashlib
 import json
 import re
+import subprocess
 import sys
 from pathlib import Path
 from typing import Any
@@ -33,7 +34,34 @@ EDGE_TYPES = {
     "parent_doc",
 }
 FRONTMATTER_BOUNDARY = "---"
-ROOT_DIR = Path(__file__).resolve().parents[3]
+
+
+def _resolve_project_root() -> Path:
+    """Locate the project root via git, not a fixed relative climb.
+
+    This script is deployed both at its source location
+    (skills/index/scripts/) and mirrored one directory deeper into
+    .claude/skills/index/scripts/ / .agents/skills/index/scripts/ -- a fixed
+    parents[N] climb from __file__ only lands on the real repo root from the
+    source location, not from a mirror. Same approach as
+    skills/j-close-story/scripts/check-story-closeable.sh,
+    check-privatized.sh:138, and compute-scope-divergence.sh:54 use for the
+    equivalent bash case.
+    """
+    try:
+        result = subprocess.run(
+            ["git", "rev-parse", "--show-toplevel"],
+            cwd=Path(__file__).resolve().parent,
+            capture_output=True,
+            text=True,
+            check=True,
+        )
+        return Path(result.stdout.strip())
+    except (subprocess.CalledProcessError, FileNotFoundError, OSError):
+        return Path.cwd()
+
+
+ROOT_DIR = _resolve_project_root()
 BOARD_ID_RE = re.compile(r"^E\d{2}(?:_S\d{2})?(?:_T\d{2})?$")
 PLAN_RE = re.compile(r"^(E\d{2}_S\d{2}(?:_T\d{2})?)-plan$")
 SUMMARY_RE = re.compile(r"^(E\d{2}_S\d{2}(?:_T\d{2})?)-summary$")

package/skills/j-brainstorm/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.brainstorm
-description: Polyfill alias of the brainstorm skill under a collision-safe directory name. Identical behavior to /brainstorm — Engage the scrum-master agent in a focused planning session to define, refine, or challenge features, improvements, tasks, stories, and epics. The agent asks probing questions, challenges assumptions, and helps shape ideas into actionable backlog items. Use when the bare /brainstorm form is shadowed by another tool's own built-in command of the same name.
+description: Engage the scrum-master agent in a focused planning session to define, refine, or challenge features, improvements, tasks, stories, and epics. The agent asks probing questions, challenges assumptions, and helps shape ideas into actionable backlog items.
 keywords:
   - brainstorm
   - plan
@@ -8,7 +8,6 @@ keywords:
   - feature planning
   - requirements
   - j-brainstorm
-  - polyfill
 examples:
   - "let's brainstorm ideas for X"
   - "I need to plan a new feature"
@@ -19,9 +18,9 @@ metadata:
 
 # Brainstorm — Collaborative Planning with the Scrum Master
 
-This skill is a literal-directory-name duplicate of `skills/brainstorm/`. It exists so that `/j-brainstorm` (and `j.j-brainstorm`) give a guaranteed-unshadowed way to reach the same flow as `/brainstorm`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/brainstorm` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-brainstorm/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-brainstorm/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh brainstorm` from `skills/brainstorm/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/brainstorm/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Instructions
 

package/skills/j-btw/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.btw
-description: Polyfill alias of the btw skill under a collision-safe directory name. Identical behavior to /btw — Capture a new mission (feature, change, or addition) and fit it into the project's Epic/Story structure, then choose to implement now or defer. Use when the bare /btw form is shadowed by another tool's own built-in command of the same name.
+description: Capture a new mission (feature, change, or addition) and fit it into the project's Epic/Story structure, then choose to implement now or defer.
 keywords:
   - btw
   - capture
@@ -8,7 +8,6 @@ keywords:
   - new idea
   - side task
   - j-btw
-  - polyfill
 examples:
   - "btw I also need to add X"
   - "capture this as a new task"
@@ -19,9 +18,9 @@ metadata:
 
 # BTW — Capture a Mission
 
-This skill is a literal-directory-name duplicate of `skills/btw/`. It exists so that `/j-btw` (and `j.j-btw`) give a guaranteed-unshadowed way to reach the same flow as `/btw`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/btw` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-btw/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-btw/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh btw` from `skills/btw/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/btw/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Instructions
 

package/skills/j-clearify/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.clearify
-description: Polyfill alias of the clearify skill under a collision-safe directory name. Identical behavior to /clearify — Clarifies ambiguous, dense, or under-specified prompts and conversation on request — inspects an attached prompt or falls back to the current conversation and surfaces plain-language clarifications with examples. Use when the bare /clearify form is shadowed by another tool's own built-in command of the same name.
+description: Clarifies ambiguous, dense, or under-specified prompts and conversation on request — inspects an attached prompt or falls back to the current conversation and surfaces plain-language clarifications with examples.
 keywords:
   - clarify this
   - clarify
@@ -10,7 +10,6 @@ keywords:
   - explain that again
   - can you simplify
   - j-clearify
-  - polyfill
 examples:
   - "clarify this"
   - "clearify the last message"
@@ -22,9 +21,9 @@ alias: wtf
 
 # Clearify — Ambiguity Clarification
 
-This skill is a literal-directory-name duplicate of `skills/clearify/`. It exists so that `/j-clearify` (and `j.j-clearify`) give a guaranteed-unshadowed way to reach the same flow as `/clearify`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/clearify` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-clearify/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-clearify/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh clearify` from `skills/clearify/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/clearify/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 > **Note on the `alias: wtf` frontmatter field:** this repo has no runtime mechanism that reads an `alias` key to route slash commands — no existing `SKILL.md` implements one, and `project/configs/workflow.json` has no alias registry. The field here documents the intended relationship only. `/wtf` is made invocable as a working alias by the companion skill folder at `skills/j-wtf/SKILL.md`, which delegates to these same instructions.
 

package/skills/j-close-story/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.close-story
-description: Polyfill alias of the close-story skill under a collision-safe directory name. Identical behavior to /close-story — Close a story by verifying all tasks are in terminal state, extracting actual diff stats per task, computing scope divergence flags, and writing closure metadata to task frontmatter. Use when the bare /close-story form is shadowed by another tool's own built-in command of the same name.
+description: Close a story by verifying all tasks are in terminal state, extracting actual diff stats per task, computing scope divergence flags, and writing closure metadata to task frontmatter.
 keywords:
   - close story
   - close
@@ -8,7 +8,6 @@ keywords:
   - complete story
   - story done
   - j-close-story
-  - polyfill
 examples:
   - "close story E17_S06"
   - "/close-story E32_S06"
@@ -18,9 +17,9 @@ examples:
 
 # close-story — Close a Story
 
-This skill is a literal-directory-name duplicate of `skills/close-story/`. It exists so that `/j-close-story` (and `j.j-close-story`) give a guaranteed-unshadowed way to reach the same flow as `/close-story`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/close-story` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-close-story/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-close-story/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh close-story` from `skills/close-story/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/close-story/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Purpose
 
@@ -45,7 +44,7 @@ This file is generated/synced by `scripts/generate-j-alias.sh close-story` from
 Run the guard script:
 
 ```bash
-bash skills/j-close-story/scripts/check-story-closeable.sh <story-id>
+bash "$([ -f skills/j-close-story/scripts/check-story-closeable.sh ] && echo skills/j-close-story/scripts/check-story-closeable.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/check-story-closeable.sh)" <story-id>
 ```
 
 - If it exits with code 0 and prints `CLOSEABLE`, continue.
@@ -72,7 +71,7 @@ For **each task ID** listed in the story's `tasks:` frontmatter array:
 
 3. **Run per-task diff stat extraction:**
    ```bash
-   bash skills/j-close-story/scripts/extract-task-diff-stats.sh <task-id>
+   bash "$([ -f skills/j-close-story/scripts/extract-task-diff-stats.sh ] && echo skills/j-close-story/scripts/extract-task-diff-stats.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/extract-task-diff-stats.sh)" <task-id>
    ```
    The script outputs:
    ```
@@ -83,11 +82,11 @@ For **each task ID** listed in the story's `tasks:` frontmatter array:
 
 4. **Write stats to task frontmatter:**
    ```bash
-   bash skills/j-close-story/scripts/update-task-frontmatter.sh \
+   bash "$([ -f skills/j-close-story/scripts/update-task-frontmatter.sh ] && echo skills/j-close-story/scripts/update-task-frontmatter.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/update-task-frontmatter.sh)" \
      "project/board/tasks/<task-id>_*.md" \
      actual_files_changed <N>
 
-   bash skills/j-close-story/scripts/update-task-frontmatter.sh \
+   bash "$([ -f skills/j-close-story/scripts/update-task-frontmatter.sh ] && echo skills/j-close-story/scripts/update-task-frontmatter.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/update-task-frontmatter.sh)" \
      "project/board/tasks/<task-id>_*.md" \
      actual_lines_delta <N>
    ```
@@ -104,13 +103,13 @@ For **each task ID** listed in the story's `tasks:` frontmatter array:
 
    Run the divergence computation:
    ```bash
-   DIVERGENCE_FLAG=$(bash skills/j-close-story/scripts/compute-scope-divergence.sh \
+   DIVERGENCE_FLAG=$(bash "$([ -f skills/j-close-story/scripts/compute-scope-divergence.sh ] && echo skills/j-close-story/scripts/compute-scope-divergence.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/compute-scope-divergence.sh)" \
      "<execution_scope>" <actual_files_changed> <actual_lines_delta>)
    ```
 
    Write the result to task frontmatter:
    ```bash
-   bash skills/j-close-story/scripts/update-task-frontmatter.sh \
+   bash "$([ -f skills/j-close-story/scripts/update-task-frontmatter.sh ] && echo skills/j-close-story/scripts/update-task-frontmatter.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/update-task-frontmatter.sh)" \
      "project/board/tasks/<task-id>_*.md" \
      scope_divergence_flag "$DIVERGENCE_FLAG"
    ```
@@ -127,7 +126,7 @@ For **each task ID** listed in the story's `tasks:` frontmatter array:
    frontmatter writes and divergence computation above, check whether this
    task's own work is permanently excluded from the public mirror:
    ```bash
-   bash skills/j-close-story/scripts/check-privatized.sh <task-id> "project/board/tasks/<task-id>_*.md"
+   bash "$([ -f skills/j-close-story/scripts/check-privatized.sh ] && echo skills/j-close-story/scripts/check-privatized.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/check-privatized.sh)" <task-id> "project/board/tasks/<task-id>_*.md"
    ```
    This is a **static** check (per `$([ -f templates/SCRUM_BOARD_SCHEMA.md ] && echo templates/SCRUM_BOARD_SCHEMA.md || echo node_modules/@jenga-ai/agent/templates/SCRUM_BOARD_SCHEMA.md)`'s "Static
    vs. Reactive Status Setting" section) — it derives the task's touched-file
@@ -177,7 +176,7 @@ against the story's own direct commits (E51_S03_T03) **before** deciding the
 story's terminal status:
 
 ```bash
-bash skills/j-close-story/scripts/check-privatized.sh <story-id> "project/board/stories/<story-id>_*.md"
+bash "$([ -f skills/j-close-story/scripts/check-privatized.sh ] && echo skills/j-close-story/scripts/check-privatized.sh || echo node_modules/@jenga-ai/agent/skills/j-close-story/scripts/check-privatized.sh)" <story-id> "project/board/stories/<story-id>_*.md"
 ```
 
 Passing the story id (`E##_S##`) rather than a task id makes the script's

package/skills/j-close-story/scripts/check-story-closeable.sh

@@ -21,11 +21,18 @@ usage() {
   exit 1
 }
 
-# Locate the project root relative to this script's location.
-# Script lives at: skills/j-close-story/scripts/check-story-closeable.sh
-# Project root is three levels up (skills/ → repo root).
+# Locate the project root via git, not a fixed relative climb — this script
+# is deployed both at its source location (skills/j-close-story/scripts/)
+# and mirrored to .claude/skills/j-close-story/scripts/, and a fixed "three
+# levels up" climb lands in the wrong place from the mirror. Same approach as
+# this directory's siblings: check-privatized.sh:138 and
+# compute-scope-divergence.sh:54.
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-PROJECT_ROOT="$(cd "$SCRIPT_DIR/../../.." && pwd)"
+PROJECT_ROOT="$(git -C "$SCRIPT_DIR" rev-parse --show-toplevel 2>/dev/null || true)"
+if [[ -z "$PROJECT_ROOT" ]]; then
+  echo "ERROR: could not locate project root (git rev-parse --show-toplevel failed from $SCRIPT_DIR)" >&2
+  exit 1
+fi
 STORIES_DIR="$PROJECT_ROOT/project/board/stories"
 TASKS_DIR="$PROJECT_ROOT/project/board/tasks"
 

package/skills/j-commit/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.commit
-description: Polyfill alias of the commit skill under a collision-safe directory name. Identical behavior to /commit — Commit implemented epic, story, or task work using the EST naming convention. Also handles user-action prerequisites and new-epic boundaries. Use after completing any EST work item. Use when the bare /commit form is shadowed by another tool's own built-in command of the same name.
+description: Commit implemented epic, story, or task work using the EST naming convention. Also handles user-action prerequisites and new-epic boundaries. Use after completing any EST work item.
 keywords:
   - commit
   - save
@@ -8,7 +8,6 @@ keywords:
   - done
   - push
   - j-commit
-  - polyfill
 examples:
   - "commit this work"
   - "save my changes"
@@ -17,9 +16,9 @@ examples:
 
 # Commit — Commit Completed Work
 
-This skill is a literal-directory-name duplicate of `skills/commit/`. It exists so that `/j-commit` (and `j.j-commit`) give a guaranteed-unshadowed way to reach the same flow as `/commit`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/commit` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-commit/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-commit/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh commit` from `skills/commit/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/commit/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Inline Mode (called by /do for inline-scoped tasks)
 

package/skills/j-continue/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.continue
-description: Polyfill alias of the continue skill under a collision-safe directory name. Identical behavior to /continue — Check project status across PROJECT_SUMMARY.md, epics, and stories to determine what should be done next. Reports "All done!" if everything is complete. Use when the bare /continue form is shadowed by another tool's own built-in command of the same name.
+description: Check project status across PROJECT_SUMMARY.md, epics, and stories to determine what should be done next. Reports "All done!" if everything is complete.
 keywords:
   - continue
   - next
@@ -8,7 +8,6 @@ keywords:
   - what's next
   - status
   - j-continue
-  - polyfill
 examples:
   - "what should I do next?"
   - "continue with the project"
@@ -17,17 +16,17 @@ examples:
 
 # Continue — Pick Up the Next Work Item
 
-This skill is a literal-directory-name duplicate of `skills/continue/`. It exists so that `/j-continue` (and `j.j-continue`) give a guaranteed-unshadowed way to reach the same flow as `/continue`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/continue` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-continue/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-continue/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh continue` from `skills/continue/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/continue/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Instructions
 
 1. **Check `project/PROJECT_SUMMARY.md`** — Determine if there is outstanding work at the project level.
 
-2. **Check `project/epics/`** — If the project summary is done, check if any epics have remaining work.
+2. **Check `project/board/epics/`** — If the project summary is done, check if any epics have remaining work.
 
-3. **Check `project/stories/`** — If epics are done, check if any stories have remaining work.
+3. **Check `project/board/stories/`** — If epics are done, check if any stories have remaining work.
 
    **Important:** Always check story status within an epic even if the epic itself is marked as done.
 

package/skills/j-deep-dive/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.deep-dive
-description: Polyfill alias of the deep-dive skill under a collision-safe directory name. Identical behavior to /deep-dive — >. Use when the bare /deep-dive form is shadowed by another tool's own built-in command of the same name.
+description: ">."
 metadata:
   prefered_agent: scrum_master
 keywords:
@@ -9,7 +9,6 @@ keywords:
   - "analyze thoroughly"
   - "think through"
   - j-deep-dive
-  - polyfill
 examples:
   - "deep dive on this idea"
   - "let's think through this properly"
@@ -20,9 +19,9 @@ examples:
 
 # /deep-dive — Deep Investigation Workflow
 
-This skill is a literal-directory-name duplicate of `skills/deep-dive/`. It exists so that `/j-deep-dive` (and `j.j-deep-dive`) give a guaranteed-unshadowed way to reach the same flow as `/deep-dive`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/deep-dive` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-deep-dive/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-deep-dive/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh deep-dive` from `skills/deep-dive/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/deep-dive/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Overview
 

package/skills/j-distribute/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.distribute
-description: Polyfill alias of the distribute skill under a collision-safe directory name. Identical behavior to /distribute — Distribute Jenga AI framework files from this private monorepo to one or more consuming projects via the local filesystem. Manages release type selection, version bumping, dry-run preview, per-target file copy, and a post-distribution git commit. Use when the bare /distribute form is shadowed by another tool's own built-in command of the same name.
+description: Distribute Jenga AI framework files from this private monorepo to one or more consuming projects via the local filesystem. Manages release type selection, version bumping, dry-run preview, per-target file copy, and a post-distribution git commit.
 keywords:
   - distribute
   - private distribution
@@ -9,7 +9,6 @@ keywords:
   - version bump
   - distribute to projects
   - j-distribute
-  - polyfill
 examples:
   - "/distribute"
   - "/distribute /path/to/consuming-project"
@@ -19,9 +18,9 @@ examples:
 
 # Distribute
 
-This skill is a literal-directory-name duplicate of `skills/distribute/`. It exists so that `/j-distribute` (and `j.j-distribute`) give a guaranteed-unshadowed way to reach the same flow as `/distribute`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/distribute` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-distribute/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-distribute/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh distribute` from `skills/distribute/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/distribute/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 Copies Jenga AI framework files from this monorepo to all active consuming projects registered in `distribute.config.json`. Manages the full version lifecycle: release type selection, `package.json` version bump, dry-run preview, per-target file copy, and a final git commit of the version bump.
 

package/skills/j-do/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.do
-description: Polyfill alias of the do skill under a collision-safe directory name. Identical behavior to /do — Execute tasks from the scrum board. Reads from project/todo.md, resolves each entry to its full scrum board context, and drives the developer agent through implementation with the correct sender object and communication contract. Loops until all selected tasks are done or the user exits. Use when the bare /do form is shadowed by another tool's own built-in command of the same name.
+description: Execute tasks from the scrum board. Reads from project/todo.md, resolves each entry to its full scrum board context, and drives the developer agent through implementation with the correct sender object and communication contract. Loops until all selected tasks are done or the user exits.
 keywords:
   - do
   - execute
@@ -8,7 +8,6 @@ keywords:
   - work on
   - build
   - j-do
-  - polyfill
 examples:
   - "implement the login feature"
   - "work on the API endpoint"
@@ -26,11 +25,10 @@ directory under a distinct name, so a host tool shipping its own same-named buil
 shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see
 `docs/skill-authoring.md`'s "Invocation Convention").
 
-> ⚠️ **Do not run `scripts/generate-j-alias.sh do` against this directory.** This file was
-> previously generated from `skills/do/SKILL.md`, and carried a banner saying so. That relationship
-> is inverted under the contract above: edits land here first, and `skills/do/` is the copy awaiting
-> deletion by `E50_S15`. Regenerating would overwrite this file from the stale bare directory.
-> CLAUDE.md states the same prohibition in general terms; this is the concrete instance of it.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is
+> nothing to run.** This file was previously generated from a bare `skills/do/SKILL.md` source;
+> `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for
+> this skill — edit it directly.
 
 ## `--trivial` Flag
 
@@ -194,7 +192,7 @@ After acquiring the epic lock and before writing the bundle manifest, scan all o
 
       1. **Extract expected files** from the task's frontmatter field `scope_rationale` and from the task's `## Description` section. Use a best-effort prose heuristic: split the text on whitespace and punctuation, then retain any token that either (a) contains a `/` character or (b) matches the pattern `*.*` (a dot surrounded by non-dot characters on both sides, e.g. `SKILL.md`, `foo.json`). Collect all retained tokens into a set called `expected_files`. This is intentionally permissive — false positives (expected files that were never actually changed) are acceptable and produce no report.
 
-      2. **Compute unexpected files**: let `actual_files` = the array stored at `task_changed_files[<task_id>]` in the bundle manifest (from step c.1). Compute `unexpected = actual_files − expected_files` (set difference: files in `actual_files` that have no match in `expected_files`). Matching is case-sensitive and exact against the relative path or the basename of the path — a token like `SKILL.md` matches any actual file whose basename is `SKILL.md` (e.g. `skills/do/SKILL.md`).
+      2. **Compute unexpected files**: let `actual_files` = the array stored at `task_changed_files[<task_id>]` in the bundle manifest (from step c.1). Compute `unexpected = actual_files − expected_files` (set difference: files in `actual_files` that have no match in `expected_files`). Matching is case-sensitive and exact against the relative path or the basename of the path — a token like `SKILL.md` matches any actual file whose basename is `SKILL.md` (e.g. `skills/j-do/SKILL.md`).
 
       3. **If `unexpected` is non-empty**:
          a. Write a Markdown conflict report to `project/queue/conflict-<task_id>.md` with the following structure:
@@ -235,14 +233,14 @@ After acquiring the epic lock and before writing the bundle manifest, scan all o
       2. Create the lock file: write the current ISO 8601 timestamp into `project/board/tasks/<task_id>_*.md.lock`.
       3. Update the task frontmatter field `status: Failed`.
       4. Delete the lock file immediately after the write completes.
-      Then invoke the rollback anchor cleanup procedure (see "### Bundle Manifest and Rollback Anchor" — failure path). After the rollback anchor completes, **release the developer concurrency slot** acquired in step 3 above: `scripts/release-concurrency-slot.sh developer <E##_S##> <orchestrator_session_id>` (per `### 4.4` step 2 below). Then **release the epic lock**: run `rm -f project/queue/epic-lock-<E##>.json`. Mark the bundle as failed and **halt** — do not execute any remaining tasks in the sequence.
+      Then invoke the rollback anchor cleanup procedure (see "### Bundle Manifest and Rollback Anchor" — failure path). After the rollback anchor completes, **release the developer concurrency slot** acquired in step 3 above: `"$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <E##_S##> <orchestrator_session_id>` (per `### 4.4` step 2 below). Then **release the epic lock**: run `rm -f project/queue/epic-lock-<E##>.json`. Mark the bundle as failed and **halt** — do not execute any remaining tasks in the sequence.
 
 5. **Post-bundle verification** (runs only if all tasks completed without failure):
    - Inspect each task in the bundle for `needs_docs: true` in its frontmatter.
    - **If any task has `needs_docs: true`**: invoke the tester agent once for the full story. Pass the story ID, the list of all task IDs in the bundle, and the shared worktree path. The tester is responsible for updating individual task statuses.
    - **If all tasks have `needs_docs: false`**: the developer agent self-verifies — reviews each task's implementation against its acceptance criteria without invoking the tester. After self-verification passes, write `status: Passed` and `date_completed: <YYYY-MM-DD>` (today's date) to each task file using the file-locking protocol (steps c.1–c.4 above). No tester invocation occurs.
 
-6. **Cleanup**: on successful bundle completion, invoke the rollback anchor success path (see "### Bundle Manifest and Rollback Anchor" — success path) to delete the bundle manifest. Then **release the developer concurrency slot** acquired in step 3 above: `scripts/release-concurrency-slot.sh developer <E##_S##> <orchestrator_session_id>` (per `### 4.4` step 2 below). Then **release the epic lock**: run `rm -f project/queue/epic-lock-<E##>.json`.
+6. **Cleanup**: on successful bundle completion, invoke the rollback anchor success path (see "### Bundle Manifest and Rollback Anchor" — success path) to delete the bundle manifest. Then **release the developer concurrency slot** acquired in step 3 above: `"$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <E##_S##> <orchestrator_session_id>` (per `### 4.4` step 2 below). Then **release the epic lock**: run `rm -f project/queue/epic-lock-<E##>.json`.
 
 ### Bundle Manifest and Rollback Anchor
 
@@ -443,7 +441,7 @@ After resolving the task context (step 4), passing override validation (step 4.1
 
 1. **Acquire a developer concurrency slot, then spawn a developer subagent.** Acquire first, per `### 4.4. Developer Concurrency Slot Enforcement` below (`<id>` = this task's id). On a full cap (`capacity_blocked` outcome), do not spawn anything — `### 4.4` already reverts this task's status to `Pending`, logs the `capacity_blocked` event, and tracks the consecutive-block count; stop here and let `/jenga` Phase 4 retry this task on a later wave once a slot frees up. On a successful acquire, spawn the subagent (Agent tool, `subagent_type: "developer"`) with the same sender object and context payload as step 5 would use, but with an explicit instruction added to the dispatch prompt: **do not create a worktree** — implement directly against the current checkout (the session's existing working tree), not an isolated `.claude/worktrees/<slug>` copy. This is the one concrete difference from the step-5 `task` path: everything else about how the subagent implements the task (reading the task file, following acceptance criteria, following repo conventions) is unchanged.
 
-2. **After the developer subagent reports implementation complete**, first **release the developer concurrency slot** acquired in step 1: `scripts/release-concurrency-slot.sh developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2) — this subagent's session has ended, so the slot is released now regardless of what it reports, before the smoke test result is even known. Then run the smoke test harness using the same invocation convention as `### 4.2. Inline Execution Path`:
+2. **After the developer subagent reports implementation complete**, first **release the developer concurrency slot** acquired in step 1: `"$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2) — this subagent's session has ended, so the slot is released now regardless of what it reports, before the smoke test result is even known. Then run the smoke test harness using the same invocation convention as `### 4.2. Inline Execution Path`:
    - Run `bash "$([ -f scripts/smoke-harness.sh ] && echo scripts/smoke-harness.sh || echo node_modules/@jenga-ai/agent/scripts/smoke-harness.sh)" <changed_file>...`, passing the paths the subagent changed. With no arguments the harness infers them from `git diff --name-only HEAD`. It exits `0` on pass and `1` on failure.
    - If neither `scripts/smoke-harness.sh` nor `node_modules/@jenga-ai/agent/scripts/smoke-harness.sh` exists, log a warning and treat the result as a pass:
      ```
@@ -468,7 +466,7 @@ This is a self-contained, reusable procedure with two current callers — `### 4
 2. **Create a worktree** for the task, named `<E##_S##_T##-short-slug>` per standard Worktree Management conventions, if one does not already exist for this task. (A task dispatched under `light` scope, or forced `inline` via `--trivial`, never had one — both premises skip worktree creation — so this step always creates a fresh worktree in that case.)
 3. **Acquire a developer concurrency slot** (per `### 4.4. Developer Concurrency Slot Enforcement` below, `<id>` = this task's id). This fallback spawn is a distinct developer-subagent lifecycle from whatever `light`/`trivial` attempt preceded it — that attempt's own slot, if any, was already acquired and released around it (`### 4.3` step 1/2, or no slot at all for a `--trivial`-forced inline attempt, which never spawns a subagent) — so this step always acquires its own fresh slot. On a full cap (`capacity_blocked` outcome), do not spawn anything here either: apply `### 4.4`'s Pending-revert/log/consecutive-block handling for this task id and stop the fallback. The task remains exactly as the reduced-overhead attempt left it (any commits already made by that attempt stay in the worktree/branch just created in step 2), and `/jenga` Phase 4 retries it on a later wave once a slot frees up.
 4. **Spawn a developer subagent** in that worktree and have it pick up from the current state of the code (the changes already made by the reduced-overhead attempt are still present in the working tree / already committed, if any commit occurred — the subagent continues from there rather than starting over).
-5. **Invoke the tester agent** per the normal `### 5. Invoke the developer agent` flow's contract — full sender object, commit SHAs, worktree path. The tester is responsible for the terminal status write, exactly as in the standard `task`-scope pipeline. Once this developer subagent's session ends — tester-verified, failed, or errored — **release the slot** acquired in step 3: `scripts/release-concurrency-slot.sh developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2).
+5. **Invoke the tester agent** per the normal `### 5. Invoke the developer agent` flow's contract — full sender object, commit SHAs, worktree path. The tester is responsible for the terminal status write, exactly as in the standard `task`-scope pipeline. Once this developer subagent's session ends — tester-verified, failed, or errored — **release the slot** acquired in step 3: `"$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2).
 6. **Emit a clear, non-fatal fallback notice** to the user/orchestrator, using the message matching the caller's origin:
    - origin `light`:
      ```
@@ -497,7 +495,7 @@ Detecting which case applies is mechanical: if the dispatch context already cont
 
 1. **Acquire.** Before spawning the developer subagent (or the bundle's shared developer subagent), run:
    ```
-   scripts/acquire-concurrency-slot.sh developer <id> <orchestrator_session_id>
+   "$([ -f scripts/acquire-concurrency-slot.sh ] && echo scripts/acquire-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/acquire-concurrency-slot.sh)" developer <id> <orchestrator_session_id>
    ```
    Per that script's own exit-code contract:
    - **Exit `0`** — slot acquired. Proceed to spawn the subagent exactly as documented at the calling site, and carry out step 2 below (release) once that subagent's session ends.
@@ -523,7 +521,7 @@ Detecting which case applies is mechanical: if the dispatch context already cont
 
 2. **Release — on every subagent exit path.** Once the spawned developer subagent's session ends — whether it reports success, failure, or an error/exception — run:
    ```
-   scripts/release-concurrency-slot.sh developer <id> <orchestrator_session_id>
+   "$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <id> <orchestrator_session_id>
    ```
    This call is unconditional and must be reached from every branch that follows a successful acquire in step 1: the success path, the ordinary `/do` failure→skip→`Pending` path, and any other error/exception branch that aborts the calling section early. `release-concurrency-slot.sh` is idempotent by its own contract (a no-op success if the holder entry is already absent), so calling it defensively — even in a code path where it is uncertain whether the acquire actually landed — is always safe and never an error.
 
@@ -564,7 +562,7 @@ The developer agent will:
 - Implement, commit at milestones, and invoke the tester agent
 - Return when the tester has verified the work
 
-**Release the slot on every exit path.** Once the developer agent's session ends — whether it returns having implemented and been verified by the tester, fails outright, or errors — release the slot immediately: `scripts/release-concurrency-slot.sh developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2). This applies uniformly to the success path above, `/jenga`'s existing "`/do` failure (background agent)" skip-to-`Pending` path (`skills/jenga/SKILL.md`'s Edge Cases), and any other error/exception that aborts this section early.
+**Release the slot on every exit path.** Once the developer agent's session ends — whether it returns having implemented and been verified by the tester, fails outright, or errors — release the slot immediately: `"$([ -f scripts/release-concurrency-slot.sh ] && echo scripts/release-concurrency-slot.sh || echo node_modules/@jenga-ai/agent/scripts/release-concurrency-slot.sh)" developer <task_id> <orchestrator_session_id>` (per `### 4.4` step 2). This applies uniformly to the success path above, `/jenga`'s existing "`/do` failure (background agent)" skip-to-`Pending` path (`skills/jenga/SKILL.md`'s Edge Cases), and any other error/exception that aborts this section early.
 
 ### 5.1. Intent-vs-Diff Check (needs_docs: false only)
 

package/skills/j-doc/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: j.doc
-description: Polyfill alias of the doc skill under a collision-safe directory name. Identical behavior to /doc — Generate or update a documentation file by resolving a target path to a clear documentation objective before writing. Use when the bare /doc form is shadowed by another tool's own built-in command of the same name.
+description: Generate or update a documentation file by resolving a target path to a clear documentation objective before writing.
 metadata:
   prefered_agent: developer
 keywords:
@@ -10,7 +10,6 @@ keywords:
   - generate docs
   - update docs
   - j-doc
-  - polyfill
 examples:
   - "/doc"
   - "/doc docs/API.md"
@@ -22,9 +21,9 @@ examples:
 
 # Doc — Documentation Synthesis and Regeneration
 
-This skill is a literal-directory-name duplicate of `skills/doc/`. It exists so that `/j-doc` (and `j.j-doc`) give a guaranteed-unshadowed way to reach the same flow as `/doc`, even if a host tool's own built-in command of the same name would otherwise shadow or override the bare `/doc` alias (Claude Code's native skill resolution is a literal-string, directory-name-based match — see `docs/skill-authoring.md`'s "Invocation Convention").
+`skills/j-doc/` is the **canonical, hand-edited** directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the `E50` reopening of 2026-09-09, which promoted `skills/j-doc/` from generated twin to sole canonical form). The `j-` prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see `docs/skill-authoring.md`'s "Invocation Convention").
 
-This file is generated/synced by `scripts/generate-j-alias.sh doc` from `skills/doc/SKILL.md` — do not hand-edit it; re-run the generator instead to pick up source changes.
+> ⚠️ **`scripts/generate-j-alias.sh` was retired by `E50_S14` and no longer exists — there is nothing to run.** This file was previously generated from a bare `skills/doc/SKILL.md` source; `E50_S15` deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.
 
 ## Input Format