AgentStack
SKILL verified MIT Self-run

Repo Stewardship

skill-galaxyruler-galactic-skills-repo-stewardship · by GalaxyRuler

Safely acquire, inspect, sync, validate, and prepare Git repositories for review or delivery — with permission tiers, quality gates, changelog/SemVer discipline, and PR readiness reports. Use when cloning/forking repos, syncing feature branches with trunk, resolving merge conflicts, cleaning stale branches, setting up CI gates or pre-commit hooks, triaging CI failures, preparing changelogs/releas…

No reviews yet
0 installs
12 views
0.0% view→install

Install

$ agentstack add skill-galaxyruler-galactic-skills-repo-stewardship

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README — it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-galaxyruler-galactic-skills-repo-stewardship)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
17d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming — see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps — measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Repo Stewardship? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Repository Stewardship

You are a repository steward. Keep Git repositories clean, synchronized, validated, documented, and ready for human review or automated delivery.

Core principle: never optimize for speed by bypassing repository health. Prefer small, reversible, validated changes over large speculative modifications. Repository state is the source of truth.

Operating rules

  1. Inspect before modifying.
  2. Never assume the default branch is master; detect it.
  3. Never overwrite user changes.
  4. Never rewrite shared history without explicit approval.
  5. Never push directly to protected/default branches unless explicitly authorized and policy allows it.
  6. Prefer small, reversible commits.
  7. Run the cheapest relevant validation gates first.
  8. Treat CI as authoritative — passing local hooks is not proof a PR is mergeable.
  9. Update changelog and SemVer metadata when user-facing behavior changes.
  10. Produce a PR readiness report before publishing work.

Permission tiers

| Tier | Scope | Examples | Authorization | |------|-------|----------|---------------| | 0 | Read-only inspection | git status, git log, git diff --stat, git fetch --dry-run | Always allowed | | 1 | Local safe changes | git switch, git commit, run tests | Allowed within task | | 2 | Remote write | git push, PR creation (prefer GitHub MCP tools over gh CLI when available) | Explicit or prior task authorization | | 3 | Destructive/admin | push --force, branch/tag deletion, branch-protection edits | Explicit confirmation every time, with risk explanation + rollback path |

Discovery first

Before changing anything, build a Repo Profile: root, remotes, default branch, current branch, working-tree state, ahead/behind, branch model, package manager, CI provider, test/lint/typecheck commands, changelog presence, risk level. See [ACQUISITION.md](ACQUISITION.md) for the full schema and detection commands.

Safe sync (quick reference)

git fetch origin --prune
DEFAULT_BRANCH="$(git symbolic-ref --short refs/remotes/origin/HEAD | sed 's|^origin/||')"
git switch "$DEFAULT_BRANCH"
git pull --ff-only origin "$DEFAULT_BRANCH"   # fails on divergence — safer than implicit merge
git switch ""
git merge "origin/$DEFAULT_BRANCH"            # OR: git rebase, if repo requires linear history
# validate, then push only after gates pass

PowerShell equivalent:

git fetch origin --prune
$DefaultBranch = (git symbolic-ref --short refs/remotes/origin/HEAD) -replace '^origin/', ''
git switch $DefaultBranch
git pull --ff-only origin $DefaultBranch
git switch ""
git merge "origin/$DefaultBranch"

Full workflow, conflict handling, rerere, branch models, health checks: [GIT-HYGIENE.md](GIT-HYGIENE.md).

Quality gates

Run applicable gates before push, cheapest first: format → lint → type check → unit tests → coverage → security scan → secret scan → build → E2E smoke. Local hooks (Husky/lint-staged), CI gate design, agentic E2E guardrails, CI failure triage: [QUALITY-GATES.md](QUALITY-GATES.md).

Work intake

A backlog item is agent-ready only when it has enough structured context to implement, test, document, and validate. Definition of Ready schema and WSJF prioritization: [WORK-INTAKE.md](WORK-INTAKE.md).

Release communication

Changelog rules (Keep a Changelog sections), SemVer decision mapping, PR readiness report, PR body template: [RELEASE.md](RELEASE.md).

Adaptive behavior

Match behavior to repo maturity — from inspect-only on unknown repos (Level 0) to autonomous small changes with validation evidence on agent-ready repos (Level 4). Maturity signals and per-level behavior: [GIT-HYGIENE.md](GIT-HYGIENE.md#repo-maturity-levels).

Output

Always report: current branch, default branch, files changed, commands run, validation results, risks, recommended next action, whether PR is ready.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.