Create Pr logo

Create Pr

Organization
HLND2T
create-pr

Create a GitHub pull request from staged task changes or an already-committed current branch. Deliver source/config/ reference changes together with their computed source-owned `bin_artifacts` closure. PR validation routing is owned by the default-branch trusted workflow; snapshots, gamedata, and manifests are Release-derived only.

Overview

PublisherHLND2T
RepositoryCS2_VibeSignatures
Skill namecreate-pr
Stars
67
Forks
11
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by HLND2T on GitHub. Read the source before you install it.

Installation

Install the Create Pr AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/HLND2T/CS2_VibeSignatures/tree/main/.claude/skills/create-pr
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/HLND2T/CS2_VibeSignatures.git /tmp/CS2_VibeSignatures
mkdir -p .claude/skills
cp -r /tmp/CS2_VibeSignatures/.claude/skills/create-pr .claude/skills/create-pr
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Create Pr in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Create Pr on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Create Pr is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Create Pull Request

Create one pull request using exactly one delivery mode:

  • staged-delivery — commit the caller's explicitly staged source-owned change set, push a dev* branch, and create the PR.
  • committed-branch — push an existing clean non-main branch that is ahead of origin/main, without rewriting or supplementing its commits.

Both modes preserve the captured source-owned change set. The PR workflow independently selects its validation path after the PR exists; this skill neither predicts nor influences that routing.

Inputs

  • gamever — optional descriptive value for the PR body. Never resolve or require it for local delivery.
  • branch — optional dev* branch name for staged-delivery. If omitted on main, derive dev-<topic> from the staged change. Ignore it in committed-branch.
  • commit_title — optional Conventional Commit title for staged-delivery.
  • pr_title / pr_body — optional PR text. Otherwise derive truthful text from the delivered diff and caller-supplied test results.
  • issue — optional issue number. Add Closes #<issue> to the PR body.

Safety Rules

  • Run from the repository root and require an origin remote plus successful gh auth status.
  • Never commit directly to main, force-push, amend existing commits, or use git add -A / git add ..
  • Preserve unrelated untracked files and never stage them.
  • Require zero unstaged tracked changes at invocation. This forbids partially staged paths and unrelated work.
  • In staged-delivery, the initial staged path list is the immutable authorized change set. It must include every affected/downstream canonical artifact under bin_artifacts/ that belongs to the source/config/reference change.
  • Reject tracked gamesymbols/**, gamedata/**, release-manifests/**, and bin/**/*.yaml; these are forbidden legacy or Release-derived namespaces, not PR deliverables.
  • In committed-branch, require an empty index, a non-main attached branch, at least one commit ahead of origin/main, and a non-empty origin/main...HEAD diff. Never create a supplemental publication commit.
  • PR validation routing, including whether isolated rebuild and C++ validation run, belongs only to the trusted default-branch workflow after the PR exists. Do not classify or predict its validation path locally.
  • Snapshot/gamedata publication runs only in the release pipeline. Never claim workflow-owned CI gates passed locally.
  • Commit and push commands may take longer than an interactive timeout. Wait for their real exit status; do not infer success from elapsed time.

Step 1: Select and Guard the Delivery Mode

Record the current state and fetch the target branch:

bash
git branch --show-current
git status --short
git diff --cached --quiet
git diff --cached --name-only
git diff --cached --name-status
git diff --cached --stat
git diff --name-only
git ls-files --others --exclude-standard
git remote
gh auth status
git fetch origin main --prune
git rev-parse HEAD
git rev-parse origin/main
git rev-list --count origin/main..HEAD
git diff --name-only origin/main...HEAD
git diff --name-status origin/main...HEAD
git diff --stat origin/main...HEAD

On Windows PowerShell, use explicit argument arrays and capture $LASTEXITCODE for native commands. Do not use the automatic $args variable as a named parameter:

powershell
function Run-Native {
    param(
        [Parameter(Mandatory)][string]$Name,
        [Parameter(Mandatory)][string]$Executable,
        [Parameter(Mandatory)][string[]]$CommandArgs
    )
    $output = & $Executable @CommandArgs 2>&1
    $exitCode = $LASTEXITCODE
    Write-Output "<<<$Name exit=$exitCode>>>"
    if ($null -ne $output) { $output | ForEach-Object { $_.ToString() } }
    Write-Output "<<<END $Name>>>"
}

git diff --name-only must be empty in both modes. Untracked files may remain, but record and preserve them.

Mode A — staged-delivery

Exit 1 from git diff --cached --quiet proves staged changes exist. Save the exact cached path list as INITIAL_STAGED_PATHS, review the cached diff for unrelated files or credentials, and derive the summary.

  • On main, validate the supplied/derived branch with git check-ref-format --branch <dev-branch> and require it not to exist locally or on origin.
  • On an existing dev* branch whose HEAD equals origin/main, use that branch.
  • On any other branch, stop and ask the caller to use main or a suitable dev* branch.

Mode B — committed-branch

Exit 0 means the index is empty. Require an attached branch other than main, a valid branch name, an ahead count greater than zero, a non-empty origin/main...HEAD diff, and a clean tracked worktree. Capture INITIAL_BRANCH, INITIAL_HEAD, AHEAD_COUNT, and INITIAL_COMMITTED_PATHS. Read the full diff and log; never rewrite those commits.

For either mode, set PR_BRANCH, then reject duplicate open PRs:

bash
gh pr list --state open --head <PR_BRANCH> --json url

Step 2: Revalidate the Authorized Index

Do not run formatters or generators in this skill. In staged-delivery, verify that the cached path list still equals INITIAL_STAGED_PATHS and that no unstaged tracked path exists:

bash
git diff --name-only
git diff --cached --quiet
git diff --cached --name-only
git diff --cached --stat
git diff --cached --check

Require a non-empty staged diff, review it again, and stage nothing else. Explicitly reject any forbidden legacy/output namespace listed in the Safety Rules. In committed-branch, apply the same path rejection to origin/main...HEAD, then require both index and tracked worktree to remain clean and HEAD == INITIAL_HEAD.

Step 3: Create the Source Commit

In staged-delivery, create the validated branch when currently on main:

bash
git switch -c <dev-branch>

Derive or validate <type>(scope): <summary>: begin with a verb, keep it at most 100 characters, and omit the final period. Commit exactly the existing staged index:

bash
git commit -m "<commit_title>" -m "Co-Authored-By: Codex <codex@openai.com>"

Verify the commit changed exactly INITIAL_STAGED_PATHS and the tracked worktree/index are clean. In committed-branch, create no commit and require HEAD == INITIAL_HEAD.

Step 4: Push and Create the Pull Request

Immediately before push, require the intended branch, clean index/worktree, a positive ahead count, and a final origin/main...HEAD path list equal to the captured source paths. In committed-branch, also require HEAD == INITIAL_HEAD.

Push without force:

bash
git push -u origin <PR_BRANCH>

Build a truthful body from the delivered diff and caller-supplied validation evidence. Do not predict the workflow's validation path or claim its gates have already passed. A suitable shape is:

markdown
## Summary
- <behavioral change>
- <supporting change>

## Validation
- implementation-specific tests: <commands/results supplied by the caller>
- repository PR validation: delegated to trusted `source-artifact-required` / `pr-validate` CI

Closes #<issue>

If gamever was supplied, it may be included as descriptive context, not as evidence of local validation. Omit the issue line when absent.

Create exactly one PR against main:

bash
gh pr create --base main --head <PR_BRANCH> --title "<pr_title>" --body "<pr_body>"

Report the branch, commit SHA, pushed remote, PR URL, and final committed paths. For committed-branch, report INITIAL_HEAD and the initial ahead count. If push succeeds but PR creation fails, report the remote branch and exact failure; do not delete the branch or create another commit.

Checklist

  • Exactly one delivery mode was selected and no unstaged tracked changes existed.
  • The local commit contains the source/config/reference change and complete bin_artifacts closure.
  • No tracked snapshot, gamedata, release-manifest, or bin/**/*.yaml path was added.
  • The PR body does not predict the workflow-owned validation path or claim unrun CI results.
  • No duplicate PR existed; the branch was pushed without force; exactly one PR was opened against main.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Create Pr AI skill do?

Create a GitHub pull request from staged task changes or an already-committed current branch. Deliver source/config/ reference changes together with their computed source-owned `bin_artifacts` closure. PR validation routing is owned by the default-branch trusted workflow; snapshots, gamedata, and manifests are Release-derived only.

Why use Create Pr on TypingMind?

Because you install it once and use it with any model. Create Pr is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Create Pr in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/HLND2T/CS2_VibeSignatures/tree/main/.claude/skills/create-pr. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Create Pr?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Create Pr?

As many as you like. As long as a model supports skills, you can use Create Pr with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Create Pr AI skill free?

Yes. It is published on GitHub by HLND2T under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇