feat(skills): add opt-in gh-stacked-pr skill

Author: dmltdev

.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
     {
       "name": "void-grimoire",
       "description": "Additive coding-agent skills: session journaling, learned-correction persistence, prompt expansion, context discipline, safety checks, lookup helpers, and workflow assists",
-      "version": "4.1.1",
+      "version": "4.3.0",
       "source": "./",
       "author": {
         "name": "Dmytro L.",

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "void-grimoire",
   "description": "Additive coding-agent skills: session journaling, learned-correction persistence, prompt expansion, context discipline, safety checks, lookup helpers, and workflow assists",
-  "version": "4.1.1",
+  "version": "4.3.0",
   "author": {
     "name": "Dmytro L.",
     "email": "dmltdev@proton.me"

.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "void-grimoire",
-  "version": "4.1.1",
+  "version": "4.3.0",
   "description": "Additive coding-agent skills: session journaling, learned-correction persistence, prompt expansion, context discipline, safety checks, lookup helpers, and workflow assists",
   "author": {
     "name": "Dmytro L.",

.omp-plugin/marketplace.json

@@ -9,7 +9,7 @@
     {
       "name": "void-grimoire",
       "description": "Additive coding-agent skills: session journaling, learned-correction persistence, prompt expansion, context discipline, safety checks, lookup helpers, and workflow assists",
-      "version": "4.1.1",
+      "version": "4.3.0",
       "source": "./",
       "author": {
         "name": "Dmytro L.",

README.md

@@ -33,7 +33,7 @@ Replace `session-summary` with any skill name from the list below.
 
 ## What You Get
 
-51 skills physically grouped across 12 domains, plus one repo-local operational skill. Pick by name:
+52 skills physically grouped across 12 domains, plus one repo-local operational skill. Pick by name:
 
 | Domain | Skills | Description |
 |--------|--------|-------------|
@@ -47,7 +47,7 @@ Replace `session-summary` with any skill name from the list below.
 | **documentation** | lookup-docs, index-docs, atlas-research, document-ideas, document-adr | Documentation lookup, indexing, research, idea capture, and architecture decisions |
 | **reasoning** | five-reasons-why, what-if | Causal analysis and bounded counterfactual exploration |
 | **tools** | using-herdr, using-codex, using-omp, omp-plugins, using-adhd, using-agent-browser, using-elevenlabs-tts, using-orx | External CLI and browser wrappers, invocation, preflight, fallbacks, and plugin operations |
-| **git** | enforce-git-safety, git-workflow, git-active-remote, git-commit, git-branch-policy, git-push, git-pr, gh-workflow, glab-workflow | Git safety, commit, push, branch policy, and GitHub or GitLab workflows |
+| **git** | enforce-git-safety, git-workflow, git-active-remote, git-commit, git-branch-policy, git-push, git-pr, gh-stacked-pr, gh-workflow, glab-workflow | Git safety, commit, push, branch policy, explicit-request stacked GitHub PRs, and GitHub or GitLab workflows |
 | **release** | enforce-release-safety | Package release and publication safety |
 
 ### Repo-local operational skill
@@ -62,6 +62,7 @@ Replace `session-summary` with any skill name from the list below.
 - **`quick-recap`** — Adds the final red/yellow/green status-line convention for finished, pending, or blocked responses.
 - **`engineering-recap`**: Reports completed engineering work as Problem, Decision, Check, and Next; use it for delivery summaries and handoffs, not generic document briefs.
 - **`document-adr`**: Creates convention-first Architecture Decision Records for accepted or proposed hard-to-reverse decisions.
+- **`gh-stacked-pr`**: Create stacked GitHub PRs only when explicitly requested. The ordinary workflow may suggest it but does not invoke it automatically.
 - **`learn-correction`** — When you correct the AI ("don't mock the DB", "always use snake_case"), persists the correction to your project's `AGENTS.md` / `CLAUDE.md` so it survives future sessions.
 - **`expand-prompt`** — Turn a terse request ("add dark mode") into a structured intent: relevant docs, learned rules, decomposed sub-tasks. Requires explicit user approval before any action.
 - **`strategic-compact`** — Suggests manual `/compact` at phase boundaries (planning -> implementing -> verifying) so context survives the next phase rather than waiting for arbitrary auto-compaction.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "void-grimoire",
-  "version": "4.1.1",
+  "version": "4.3.0",
   "description": "Utility skills for coding agents",
   "private": true,
   "license": "MIT",

skills-lock.json

@@ -76,6 +76,11 @@
       "sourceType": "local",
       "computedHash": "aa2de871d33d691cee3957e2bc9dcd229aea2e6557a14b76f8f84b69bc40abbf"
     },
+    "gh-stacked-pr": {
+      "source": ".",
+      "sourceType": "local",
+      "computedHash": "255dc346595f35a2f15e6e10745db3e1c03d1f9be0bdb61509a38afbc9a0af61"
+    },
     "gh-workflow": {
       "source": ".",
       "sourceType": "local",
@@ -99,7 +104,7 @@
     "git-pr": {
       "source": ".",
       "sourceType": "local",
-      "computedHash": "93b1bb63a7fadd0f922baff7828e609817b61224d28533c994c8a622ec38922a"
+      "computedHash": "47b77e46dd418375f4bf3911a9010bc0e09da674186a089bea5fcfa1a6f9507d"
     },
     "git-push": {
       "source": ".",
@@ -109,7 +114,7 @@
     "git-workflow": {
       "source": ".",
       "sourceType": "local",
-      "computedHash": "0361340edc70c88efc01907c41cd68badf72479e634a962bb023de8fcfff4df7"
+      "computedHash": "e86e22958af854d3258b1adfc3be6a5eb6ba128c13bc15d758f08a3ee873dc17"
     },
     "glab-workflow": {
       "source": ".",

skills/git/gh-stacked-pr/SKILL.md

@@ -0,0 +1,35 @@
+---
+name: gh-stacked-pr
+description: Use when the user explicitly asks to create a stacked PR or a stack of GitHub pull requests.
+---
+
+# GitHub Stacked PRs
+
+Create a chain of PRs where each PR reviews only its own layer.
+
+## When to use
+
+Only after an explicit request to create stacked PRs.
+The common workflow may suggest stacking, but does not invoke
+this skill automatically.
+
+## Workflow
+
+1. Identify the repository, root base branch, and layer order.
+   Reuse existing branches and PRs where appropriate.
+2. Ensure the layer branches are published. If publishing is
+   needed, obtain push authorization rather than pushing implicitly.
+3. Use `gh-workflow` for GitHub operations and `git-pr` for titles
+   and bodies. Create missing PRs bottom-to-top:
+   - First PR targets the root base.
+   - Each subsequent PR targets the preceding layer's branch.
+4. Add links to the other PRs in each body, showing stack order
+   and dependencies.
+
+## Result
+
+Return the PR links in stack order, with each source and base branch.
+State the bottom-to-top merge order; do not merge automatically.
+
+Creating a stack does not authorize commits, history rewrites,
+force pushes, or branch deletion.

skills/git/git-pr/SKILL.md

@@ -17,6 +17,10 @@ Create or update reviewer-facing PR/MR text only when explicitly requested. Resp
 | Procedure | Resolve active remote, choose host adapter, inspect diff/range, draft or apply PR/MR text. |
 | Proof | Report PR/MR URL or rendered title/body preview; no unrequested push/commit occurred. |
 
+## Stacked creation is opt-in
+
+When the user explicitly asks to create stacked GitHub PRs, use `gh-stacked-pr` for topology and creation order; this skill still owns each layer's title/body. An ordinary PR request does not load that skill or create a stack. Suggesting stacked reviews is allowed, but wait for an explicit creation request.
+
 ## Preconditions
 
 - PR/MR intent is explicit.

skills/git/git-workflow/SKILL.md

@@ -23,6 +23,7 @@ Route git work by explicit user intent. A later git action is never implied by a
 |---|---|---|
 | "commit this", "make a commit", "write/commit these changes" | `git-commit` | push, PR/MR |
 | "push this", "publish branch", "update remote" | `git-push` | commit, PR/MR |
+| "create stacked PRs", "create a stacked GitHub PR", "create a stack of GitHub pull requests" | `gh-stacked-pr` with `git-pr` for each layer | unrequested commit, push, branch/history changes, merge |
 | "open/create/update PR", "open/create/update MR", "draft PR description" | `git-pr` | commit, push |
 | "commit and push" | `git-commit` then `git-push` | PR/MR |
 | "commit, push, and open PR/MR" | `git-commit`, `git-push`, `git-pr` | nothing beyond named actions |
@@ -38,6 +39,8 @@ Route git work by explicit user intent. A later git action is never implied by a
 - Do not infer push from a successful commit.
 - Do not ask when the user already named the action sequence.
 - If a mutating next step is ambiguous, choose the narrower safe action and stop.
+- GitHub stacked PR creation is opt-in. Load `gh-stacked-pr` only after the user explicitly asks to create a stack, not because changes are dependent or a branch already belongs to one.
+- The ordinary workflow may suggest stacking when separate dependent reviews would help; wait for the user's explicit creation request before loading or executing the stack skill.
 
 ## Required companion contracts