hashgraph-online/awesome-codex-plugins / plugins/alpoxdev/hypercore/skills/hyper-git
hyper-git
Verified[Hyper] Use this skill when the user asks for Git or GitHub repository work carried to a verified result: committing and pushing changes (including current-session or all changes and `&&`-separated branch propagation), creating, entering, listing, removing, or repairing Git worktrees, reading, filing, commenting on, closing, or branching from GitHub issues, or choosing a branching model, opening a pull request, and merging branches with a push. It also answers legacy `/git-maker` and `/git-worktree` invocations. Do not use it for local-only commits with no push, history rewriting, release tagging, CI authoring, or container or VM isolation.
rules/speed-and-automation.md
View on GitHub# Speed And Automation
Use this rule when the commit mode of `hyper-git` feels slow or the user explicitly asks for a faster commit-and-push path.
## Fast Path
Prefer this path before the verbose manual workflow:
```bash
scripts/git-fast.sh inspect . --jobs 4
```
Then commit each logical group with the existing commit helper:
```bash
scripts/git-commit.sh "<type>(<scope>): <imperative summary>" path/to/file1 path/to/file2
scripts/git-commit.sh --repo path/to/repo "<type>(<scope>): <imperative summary>" path/to/file1
```
After all commits succeed, reuse the inspected repo list instead of rediscovering:
```bash
scripts/git-fast.sh push /absolute/repo/path
scripts/git-fast.sh push --force /absolute/repo/path
```
If the fast helper fails or omits needed detail, fall back to `repo-discover.sh`, `repo-status.sh`, `git-commit.sh`, and `git-push.sh`.
## What The Fast Helper Automates
| Step | Automation |
|------|------|
| repository discovery | current repo/worktree short-circuit; descendant discovery prunes heavy generated directories and recognizes linked worktree `.git` files |
| multi-repo inspection | parallel per-repo status blocks via `--jobs` / `HYPER_GIT_JOBS` |
| file inventory | captures one porcelain status snapshot per repo and derives staged, unstaged, and untracked lists from it instead of running three duplicate file scans |
| legacy helpers | `repo-discover.sh` prunes generated trees, `repo-status.sh` reuses captured status/stat output, and `git-push.sh` independently preserves the same safe-push policy |
| worktree context | emits `worktree|primary` or `worktree|linked` plus git dir metadata so operators keep checkout roots separate from the common git dir |
| push | accepts explicit repo paths, uses `GIT_TERMINAL_PROMPT=0`, skips detached HEAD, blocks protected force pushes |
## Parallelization Rules
- Parallelize read-only inspection across repositories.
- Do not parallelize commits inside the same repository; staging/index state is shared.
- Repositories may be committed independently, but only when file groups and repo boundaries are already clear.
- Push only after all intended commits have succeeded.
## Slow Path Triggers
Use the slower full workflow when:
- partial hunks require `git add -p`
- staged and unstaged changes are intentionally mixed
- hooks fail and require diagnosis
- filter arguments do not clearly map to files
- a push failure needs remote/auth investigation
## Timing Expectations
The helper is meant to remove avoidable overhead, not bypass safety. It speeds up:
- duplicated repository discovery between commit and push
- large descendant tree scans by pruning common generated folders
- linked worktree runs by resolving the checkout root without assuming `.git` is a directory
- multiple-repository status checks by parallelizing read-only work
- single-repository inspection by reusing one status snapshot for both raw status and grouped file inventory
- commit-message validation in Bash without spawning `grep`
It does not remove the need to inspect diffs, group logical changes, write correct commit messages, or handle hook failures.
## Sources
> No external sources were used. Content checked 2026-09-22.
This rule set is authored in this package from repository practice. It makes no external claim, so no external source is cited.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 470k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |