Agent skill

Sync Template

by mugnavo in mugnavo/cove-monorepo

Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration.

UnlicenseAuto-check passedDevelopment

Install Sync Template

skills CLI
$ npx skills add mugnavo/cove-monorepo --skill sync-template -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install mugnavo/cove-monorepo sync-template --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/mugnavo/cove-monorepo.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/sync-template .claude/skills/sync-template && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
sync-template
GitHub stars
166
Token cost
~3.1k tokens
SKILL.md length
1,519 words
Files
1
Skills in repo
1
Repo updated
First seen
Licence
Unlicense

At a glance

Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration.

  • Cove Stack upgrades
  • SKILL.md covers Establish the project state, Resolve the upstream target, Establish the comparison base and Classify the changes, plus 3 more sections
  • Calls git and rsync; reaches github.com
  • Upstream-change reviews

What it does

Sync Template is an agent skill from mugnavo/cove-monorepo. Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration. Use for Cove Stack upgrades, starter syncs, or upstream-change reviews.

Its SKILL.md is about 3.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development. The repository describes itself as: Vite+ monorepo template with 🏝️ TanStack Start, Better Auth, Drizzle ORM, shadcn/ui. The licence is Unlicense.

When your agent uses it

  • Cove Stack upgrades
  • Upstream-change reviews

Example prompts

  • “/sync-template”

What it can do on your machine

Read from SKILL.md and the folder at commit 62114dc. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git
    • rsync

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Hosts in commands or code, which the agent is likely to contact:

    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Sync Template loads about 3.1k tokens when it runs. Until then it costs about 56 tokens; SKILL.md has 1,519 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~56
When it runs · the whole SKILL.md, loaded when a task matches
~3.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from mugnavo/cove-monorepo at commit 62114dc, republished under its Unlicense licence (© mugnavo). 1,519 words, ~3,076 tokens.

Download SKILL.mdSave it as .claude/skills/sync-template/SKILL.md (or your agent's skills folder).
name
sync-template
description
Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration. Use for Cove Stack upgrades, starter syncs, or upstream-change reviews.

Sync a project with Cove Stack

Integrate relevant changes from the project's Cove Stack template while preserving the project's own behavior.

For a comparison-only request, inspect and report without applying changes. For a sync request, prepare and validate a reviewable update. Follow existing user authorization for commits, pushes, pull requests, and deployment.

Establish the project state

  • Confirm the destination repository and its Cove Stack provenance.

  • Read .cove.jsonc when present. It may contain:

    jsonc
    {
      // Used by the sync-template skill to track this project's
      // Cove Stack source and applied template revision.
      "source": "https://github.com/mugnavo/cove",
      "revision": "<exact upstream commit>",
      "createdAt": "<ISO 8601 creation time>",
      "lastSyncedAt": "<ISO 8601 successful sync time>"
    }

    source identifies the template repository. revision is initially the downloaded commit and later the newest upstream commit fully reconciled with the project. createdAt is immutable and may be absent. lastSyncedAt is absent until the first completed sync.

  • Read its current AGENTS.md, relevant project guidance, package scripts, environment schema, and previous sync records.

  • Record the starting commit (PROJECT_START), branch, staged and unstaged changes, and untracked files.

  • Use a dedicated sync branch following the repository's naming conventions.

  • Preserve unrelated work. With a dirty checkout, prefer an isolated worktree from the relevant committed state and explain which uncommitted changes it excludes. If the update depends on those changes, resolve that dependency before integrating.

  • Do not silently stash, commit, discard, or copy secret-bearing files to obtain a clean checkout.

  • If a merge or rebase is already in progress, identify its purpose before modifying that state.

  • Establish relevant baseline check results so existing failures can be distinguished from regressions.

Resolve the upstream target

Resolve TEMPLATE_SOURCE from an explicit user choice, .cove.jsonc, previous sync records, or reliable generation and history evidence. Supported Cove Stack sources are:

  • https://github.com/mugnavo/cove
  • https://github.com/mugnavo/cove-monorepo

Workspace structure is supporting evidence, not proof of provenance. Ask rather than guess when the source remains ambiguous. Do not rewrite a recorded source unless correcting it is part of the requested work and the evidence is clear.

Discover the canonical repository's default branch:

sh
git ls-remote --symref "$TEMPLATE_SOURCE" HEAD

An explicitly requested upstream branch, tag, or commit takes precedence.

Verify a remote's URL before using it. Do not assume a remote named upstream points to TEMPLATE_SOURCE, and do not repoint an existing remote or the project's origin.

Fetch the verified ref and immediately resolve its immutable commit as TARGET. Record the source URL, ref, SHA, and fetch date. If using FETCH_HEAD, resolve it before another fetch replaces it.

If remote verification is unavailable, identify any inspected local revision as cached. Do not claim it is the latest upstream version.

Establish the comparison base

Projects created from a GitHub template or starter CLI may not share Git history with Cove Stack. Check the actual history rather than assuming the project is a fork.

Look for:

  • A real Git merge base.
  • A previously recorded upstream revision, including .cove.jsonc's revision.
  • Generation metadata or an identifiable initial starter snapshot.
  • Earlier sync commits and records of intentionally omitted changes.

Before using .cove.jsonc's revision as BASE, verify that it identifies a commit from source. Treat createdAt and lastSyncedAt only as context; timestamps and dependency versions do not establish a reliable base. Check for shallow or incomplete history before concluding that no common ancestor exists.

When a verified comparison base (BASE) exists, inspect the upstream delta:

sh
git log --oneline "$BASE..$TARGET"
git diff --stat "$BASE" "$TARGET"
git diff --name-status -M "$BASE" "$TARGET"

Review renames, deletions, and the intent of relevant commits as well as file additions.

If no reliable base exists, compare the initial project snapshot and current implementation with the target by concern. Document the uncertainty. Do not invent a base or manufacture shared history.

If the target is already integrated, review previously omitted changes before reporting that no update is needed. An older requested target requires an explicit rollback strategy; merging it does not undo newer changes.

Classify the changes

Determine ownership from file contents and local history, not directory names alone.

ConcernTreatment
Application routes, features, content, branding, and assetsPreserve project behavior; adapt only where upstream changes require it.
Framework setup, routing infrastructure, server boundaries, and shared utilitiesIntegrate relevant upstream fixes with local adaptations.
Authentication and authorizationPreserve the project's providers, roles, access rules, account flows, and optional-feature behavior while adapting API changes.
Database schema and migrationsPreserve application tables, fields, data, and applied migration history.
UI components and stylesReview local modifications before accepting upstream changes. Preserve the application's design and accessibility behavior.
Dependencies, scripts, workspace configuration, and lockfilesReconcile compatible changes together, retaining project-specific requirements.
Environment and deployment configurationPreserve deployment identities, services, startup behavior, and environment contracts.
Instructions and documentationIncorporate relevant guidance without replacing project-specific instructions with starter defaults.

Current user instructions and project guidance take precedence over historical examples.

Integrate the update

Choose the strategy that fits the actual history and previous sync process.

Shared history

On a clean, dedicated sync branch, prepare a normal merge:

sh
git merge --no-ff --no-commit "$TARGET"

--no-ff keeps a fast-forward update reviewable before committing.

Resolve conflicts by intent. Inspect files that merged cleanly too; an automatic merge can still overwrite application behavior or introduce unwanted defaults.

Continue an established selective-sync process when that is more appropriate than merging the entire upstream tree.

Unrelated history or selective updates

Apply the upstream delta from the verified base to the target rather than replacing the project with the target snapshot.

For reviewed paths, a patch generated with git diff --binary --full-index can support git apply --3way when the required base blobs are available and the index/worktree are clean. Handle moved files, deletions, and semantic changes deliberately.

When the base is unknown, reconcile relevant changes manually by concern.

Do not use a blanket copy, rsync --delete, hard reset, automatic “ours/theirs” resolution, or --allow-unrelated-histories as a shortcut. Do not create an ancestry-only merge that represents unapplied changes as integrated.

Show full SKILL.md (609 more words)Show less
Preserve adaptations during either strategy
  • Apply upstream fixes to the project's current file locations. Do not recreate old directories or parallel implementations when the project has reorganized starter code.
  • Preserve existing feature choices. New upstream functionality does not automatically belong in the application.
  • Keep project-specific dependencies and scripts. Resolve related framework versions, workspace catalogs, overrides, and peer requirements together.
  • Do not downgrade a newer local dependency solely to match upstream. Check compatibility and document intentional divergence.
  • Keep broad dependency upgrades separate unless the user requested them.
  • Review lifecycle scripts before installing dependencies. Preserve required preparation, code generation, and container build behavior.
  • Reconcile manifests first, then regenerate the lockfile with the pinned package manager. Do not replace the lockfile wholesale.
  • Regenerate route trees, environment types, and affected schemas through the project's tools. Review the resulting diffs.
  • Preserve applied migration history. Generate incremental migrations when necessary and validate them against a disposable local database.
  • Use the environment schema and sanitized examples to understand configuration. Never read local secret-bearing env files or invent credentials to make validation pass.

Validate the result

Use the destination project's current commands and testing guidance.

  • Install dependencies with the pinned toolchain after reviewing manifest and lifecycle changes.
  • Run lint/type checks and the narrowest relevant tests.
  • In projects retaining Cove Stack's Vite+ scripts, vpr lint covers linting and type checking, vpr test runs unit tests, and vpr test:e2e runs browser tests. Verify the current scripts before relying on these names.
  • If browser behavior changed, run the affected browser tests. When their configuration owns the production build and server lifecycle, do not run a duplicate build first.
  • Validate environment configuration using the project's current tooling without exposing values.
  • Check affected routes, authentication flows, navigation, forms, metadata, and responsive behavior against the baseline.
  • Exercise supported optional-feature states when their integration changes.
  • Follow the project's test-port convention. Do not stop unrelated servers to free a port.
  • Use disposable local resources for database and integration checks. Do not trigger real external side effects or run production migrations during validation.
  • For deployment changes, inspect and verify the affected build/startup path. Entrypoints may apply migrations.

Review the complete result against PROJECT_START, including staged, unstaged, and newly introduced files:

sh
git diff --check
git diff --cached --check
git ls-files -u

Check for unresolved conflicts, duplicated implementations, restored demo content, overwritten configuration, and unexpected generated files.

Separate pre-existing failures, introduced regressions, and checks blocked by the environment. Resolve introduced failures before calling the update ready.

Record and deliver

Update an existing sync record, or create docs/upstream-sync.md if none exists. Record:

  • Canonical upstream URL, ref, immutable target SHA, and date.
  • Project starting commit, comparison base, and evidence for that base.
  • Integration strategy and adopted changes.
  • Preserved customizations, intentional deviations, and outstanding omissions.
  • Validation results and remaining actions.
  • Whether the update is prepared, committed, or partial.

Keep previous entries so future syncs can reconsider omissions.

After every upstream change through TARGET has been adopted, adapted, or explicitly recorded as not applicable, and introduced regressions are resolved, update .cove.jsonc as part of the prepared sync:

  • Set revision to TARGET.
  • Set lastSyncedAt to the current ISO 8601 timestamp.
  • Preserve source, createdAt, unknown fields, and comments.

Do not update these fields for comparison-only work, a target with deferred or unresolved changes, or a failed validation. Do not invent a missing createdAt or change it during a sync.

A recorded target is not proof that all changes were integrated. Do not advance revision past unresolved work or claim a completed merge before its merge commit exists.

Deliver the branch or worktree, exact upstream target, practical changes, validation results, and remaining work. A dependency-only update is not a full starter sync. A local sync request does not itself authorize publishing, production migrations, or deployment.

© mugnavo, Unlicense. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/sync-template of mugnavo/cove-monorepo.

Open the folder on GitHubat commit 62114dc

Compare with similar skills

Sync Template next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Sync Template compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Sync Template this skillmugnavo/cove-monorepo166—~3.1kAutomated safety check: PassUnlicense
Vercel Composition Patternssupabase/supabase111k58 repos~726Automated safety check: PassMIT
Finishing a Development Branchobra/superpowers297k5 repos~1.9kAutomated safety check: PassMIT
Typescript Advanced Typesrolling-scopes/rsschool-app10k25 repos~4.2kAutomated safety check: PassMPL-2.0
PR Babysitteropeninterpreter/openinterpreter69k3 repos~4.2kAutomated safety check: PassApache-2.0
Code Review ChecklistshareAI-lab/learn-claude-code78k4 repos~1.1kAutomated safety check: PassMIT

Similar skills

  • Official

    React composition patterns that scale. An agent skill from supabase/supabase.

    111k GitHub starsUsed in 58 repos~726 tokens
    DevelopmentAuto-check passed
  • Walks the last step of a branch: confirm tests pass, detect the git environment, ask how to integrate, carry out your choice and clean up the worktree.

    297k GitHub starsUsed in 5 repos~1.9k tokens
    DevelopmentAuto-check passed
  • Typescript Advanced Types

    rolling-scopes/rsschool-app

    Master TypeScript's advanced type system including generics, conditional types, mapped types, template literals, and utility types for building type-safe applications.

    10k GitHub starsUsed in 25 repos~4.2k tokens
    DevelopmentAuto-check passed
  • PR Babysitter

    openinterpreter/openinterpreter

    Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.

    69k GitHub starsUsed in 3 repos~4.2k tokens
    DevelopmentAuto-check passed
  • Code Review Checklist

    shareAI-lab/learn-claude-code

    Reviews code against a five-part checklist covering security, correctness, performance, maintainability and testing, and reports findings in a fixed format.

    78k GitHub starsUsed in 4 repos~1.1k tokens
    DevelopmentAuto-check passed
  • Greploop

    onyx-dot-app/onyx

    Iteratively improves a PR (GitHub), MR (GitLab), or shelved changelist (Perforce) until Greptile gives it a 5/5 confidence score with zero unresolved comments.

    32k GitHub starsUsed in 4 repos~3.3k tokens
    DevelopmentAuto-check passed

Categories

Questions about Sync Template

What does Sync Template do?

Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration. Sync Template is an agent skill from mugnavo/cove-monorepo. Compare or sync projects with upstream changes from the Cove Stack template while preserving project-specific behavior and configuration.

When should I use Sync Template?

Sync Template fits situations like: cove Stack upgrades; upstream-change reviews.

How do I install Sync Template in Claude Code?

Run `npx skills add mugnavo/cove-monorepo --skill sync-template -a claude-code`. Or copy the skill folder (.agents/skills/sync-template in mugnavo/cove-monorepo) into .claude/skills/sync-template in your project. Claude Code loads it when a task matches its description.

How do I install Sync Template in Codex?

Run `npx skills add mugnavo/cove-monorepo --skill sync-template -a codex`. Or copy the skill folder (.agents/skills/sync-template in mugnavo/cove-monorepo) into .agents/skills/sync-template in your project. Codex loads it when a task matches its description.

Can I use Sync Template in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add mugnavo/cove-monorepo --skill sync-template -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/sync-template, .gemini/skills/sync-template, .github/skills/sync-template and .opencode/skills/sync-template in your project.

What does Sync Template need to run?

Going by SKILL.md and its folder, Sync Template needs the command-line tools its instructions call (git and rsync).

Does Sync Template access the network?

SKILL.md names 1 domain. In commands or code: github.com; the agent is likely to contact it when it follows the instructions. This is read from the text; nothing was executed.

Is Sync Template safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Sync Template use?

Sync Template is published under the Unlicense licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Sync Template use?

About 3.1k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Sync Template?

Skills that share tags, products or a category with Sync Template: Vercel Composition Patterns (supabase/supabase, 111k stars), Finishing a Development Branch (obra/superpowers, 297k stars), Typescript Advanced Types (rolling-scopes/rsschool-app, 10k stars) and PR Babysitter (openinterpreter/openinterpreter, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Sync Template?

mugnavo (a GitHub organization) maintains it in mugnavo/cove-monorepo, which has 166 GitHub stars. The repository was last updated on October 8, 2026.

Source: mugnavo/cove-monorepo on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.