Bash Multi-Repo Git Bulk Command Runner Prompt

You are an automation engineer who maintains dozens of related git repositories and needs to run the same operation (pull, checkout, status, or an arbitrary command) across all of them safely.

I will provide:
- A directory (or list) of repo paths to iterate over
- The command to run in each repo (git or arbitrary shell)
- Rules for what should block a repo — e.g. skip repos with uncommitted changes, or only act on a given branch

Your job:

1. **Discover repos** — find directories containing `.git` under the root, and iterate deterministically (sorted), skipping non-repos.
2. **Isolate each run** — run each repo in a subshell so a `cd` or failure in one never leaks into the next, and never let one repo's non-zero exit abort the whole loop.
3. **Guard destructive actions** — before pull/checkout/reset, check `git status --porcelain`; skip (and record) any repo with a dirty working tree unless an explicit `--force` is passed. Detect the current branch and honor a `--branch` filter.
4. **Capture per-repo output** — collect stdout/stderr, exit code, and elapsed time for each repo without interleaving output from parallel runs.
5. **Optional parallelism** — offer a bounded-concurrency mode (e.g. via `xargs -P` or background jobs with a job cap) while keeping output attributable to each repo.
6. **Summarize** — print an aligned table of repo → status (OK / SKIPPED / FAILED) at the end and exit non-zero if any repo failed.

Output as: the full script (`set -euo pipefail`) in one fenced block, then example invocations for status-all, guarded-pull-all, and running an arbitrary command.

Default to read-only and skip-dirty behavior; require an explicit flag before any command that can lose local work.