agentsclimarketplace

Redeploy frontend

Skill timi-ty/agent-forge/skills/redeploy-frontend

Trigger a Vercel redeploy of the frontend by pushing a harmless comment change. Use when the user says "redeploy", "redeploy frontend", "trigger vercel deploy", or "trigger redeploy".From its SKILL.md

Install
npx -y skills add timi-ty/agent-forge --skill redeploy-frontend

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

7.3 KB, ~1.9k tokens by cl100k_base, as published. Nobody here has run it

Redeploy Frontend

Push a trivial timestamp comment change to the current deploy branch to trigger a Vercel redeploy. Uses a git worktree so the user's working directory is never touched -- works even with uncommitted local changes. Auto-detects the package manager, deploy branch, and target file. Auto-fixes prettier formatting issues before pushing.

Workflow

Step 0: Discover project config

Resolve the repo root and detect project settings before proceeding.

Repo root:

git rev-parse --show-toplevel

All subsequent commands run from this directory. Set $REPO_ROOT to this path.

Repo name -- derive $REPO_NAME from the directory name or git remote.

Package manager -- check for lock files at repo root:

  • pnpm-lock.yaml → pnpm
  • yarn.lock → yarn
  • bun.lockb → bun
  • package-lock.json → npm

If multiple lock files exist, prefer in the order listed above.

Deploy branch -- use the currently checked-out branch:

git branch --show-current

The user is responsible for being on the correct deploy branch before invoking this skill.

Timestamp file -- find the file to inject the redeploy comment into, in priority order:

  1. Use your file-search tools (Grep) to find any .ts or .tsx file under src/ or app/ that contains the line // redeploy:. If found, that file is $TIMESTAMP_FILE.
  2. Try common entry points in order: src/app/layout.tsx, app/layout.tsx, src/main.tsx, src/App.tsx, src/index.ts, src/index.tsx
  3. If none of the above exist, ask the user: "Which file should I use to inject the redeploy timestamp comment?"

Set $BRANCH, $PKG_MANAGER, and $TIMESTAMP_FILE for use in all steps below. $TIMESTAMP_FILE is a path relative to the repo root.


Step 1: Create a worktree for the deploy branch

Fetch the latest remote state and create a clean worktree at origin/$BRANCH:

git fetch origin
git worktree add ../$REPO_NAME-wt-redeploy origin/$BRANCH

Set $DEPLOY_DIR to the absolute path of the created worktree.

If the worktree path already exists (from a previous interrupted redeploy), remove it first:

git worktree remove ../$REPO_NAME-wt-redeploy --force

Handle unpushed local commits

Check if the user has local commits on $BRANCH that are not yet on the remote:

git rev-list --count origin/$BRANCH..HEAD

If ahead > 0 and the user is on $BRANCH, ask: "You have {N} unpushed commit(s) on {branch} not yet on the remote. Include them in the redeploy?"

If yes, capture the local tip and merge it into the worktree's detached HEAD:

LOCAL_TIP=$(git rev-parse $BRANCH)
git -C $DEPLOY_DIR merge $LOCAL_TIP --no-edit

Do NOT use git -C $DEPLOY_DIR checkout $BRANCH -- git forbids the same branch in two worktrees simultaneously, and the user's main worktree already has $BRANCH checked out. The worktree is on a detached HEAD (from origin/$BRANCH), so a merge incorporates the local commits without violating the one-branch-per-tree rule.

If no, continue with the worktree at origin/$BRANCH.

Sync build-essential non-tracked files

Copy non-version-controlled files needed for builds from the main repo to the worktree. These files are gitignored and therefore absent from a fresh worktree.

for f in .env .env.local .env.production .env.development .env.production.local .env.development.local .env.test .env.test.local; do
  [ -f "$REPO_ROOT/$f" ] && cp "$REPO_ROOT/$f" "$DEPLOY_DIR/$f"
done

For monorepos, also check subdirectories -- find all .env* files (excluding node_modules and .git) in $REPO_ROOT and copy them to the same relative paths in $DEPLOY_DIR:

find "$REPO_ROOT" -mindepth 2 -name '.env*' -not -path '*/node_modules/*' -not -path '*/.git/*' | while read src; do
  rel="${src#$REPO_ROOT/}"
  mkdir -p "$DEPLOY_DIR/$(dirname "$rel")"
  cp "$src" "$DEPLOY_DIR/$rel"
done

If the project has a .env.example or .env.template in the repo root but no .env, warn the user: "No .env file found. The build may fail if environment variables are required."

Do not log or display the contents of these files -- they may contain secrets.

Step 2: Install dependencies

Install dependencies in the worktree so the build step works:

cd $DEPLOY_DIR
$PKG_MANAGER install

Optimization: For large projects where install is slow, you may symlink node_modules from the main worktree instead: ln -s $REPO_ROOT/node_modules $DEPLOY_DIR/node_modules (Unix) or mklink /J $DEPLOY_DIR\node_modules $REPO_ROOT\node_modules (Windows). This avoids a full install. Only do this if both worktrees are on the same branch and dependency versions match.

Step 3: Local pre-build check and auto-fix

All commands in this step run in $DEPLOY_DIR.

Run the build locally to catch failures before pushing:

cd $DEPLOY_DIR
$PKG_MANAGER run build

If the build passes, continue to Step 4.

If the build fails, inspect the output:

  • Prettier-only failure -- the output contains [warn] lines listing files and ends with Code style issues found. The build phase never ran. Auto-fix:

    cd $DEPLOY_DIR
    $PKG_MANAGER run format
    

    Then re-run $PKG_MANAGER run build to confirm the fix.

    • If the re-run passes, continue to Step 4. The formatted files will be staged alongside the timestamp change in Step 5.
    • If the re-run fails with a different error, treat as non-trivial (see below).
  • Non-trivial failure (ESLint errors, TypeScript type errors, build errors) -- abort, clean up the worktree (Step 7), and report the full build output to the user. Ask follow-up questions about how to resolve before pushing.

Step 4: Update the redeploy timestamp

Open $DEPLOY_DIR/$TIMESTAMP_FILE. Look for an existing line matching the pattern // redeploy:.

  • If the line exists, replace it with a new UTC timestamp: // redeploy: <timestamp>
  • If no such line exists, insert // redeploy: <timestamp> as the very first line of the file.

Generate the timestamp:

date -u +%Y-%m-%dT%H:%M:%SZ                                    # macOS/Linux/Git Bash
(Get-Date).ToUniversalTime().ToString("yyyy-MM-ddTHH:mm:ssZ")   # Windows PowerShell

Example result:

// redeploy: 2026-03-05T14:23:00Z
// ... rest of file

Step 5: Commit and push

All commands in this step run in $DEPLOY_DIR.

If prettier auto-fixed files in Step 3, stage everything and use the format-aware message:

cd $DEPLOY_DIR
git add .
git commit -m "chore: format and trigger redeploy"
git push origin $BRANCH

If no prettier fixes were needed (only the timestamp changed), stage just the timestamp file:

cd $DEPLOY_DIR
git add $TIMESTAMP_FILE
git commit -m "chore: trigger redeploy"
git push origin $BRANCH

Step 6: Confirm

Report the push result to the user. Include the commit hash from the output.

Step 7: Cleanup worktree

Remove the worktree:

git worktree remove ../$REPO_NAME-wt-redeploy

If removal fails, force it:

git worktree remove --force ../$REPO_NAME-wt-redeploy

This step runs after Step 6 on success, or after a build failure abort in Step 3.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 611,811. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.