Portable, cross-harness Agent Skills. Source and docs: github.com/marroccofella/skills
◈ MOMM · ✈ My Travel · ↗ myrepo · ♫ myvoice · ⟳ myautoness · ✓ myskills · ☕ yorky
My Travel — complete journeys and verified bookings · momm — local multi-CLI code review with a reproduction gate · myvoice — consented local voice cloning · myrepo — publish to GitHub with a live page · myautoness — self-playing by search and verified replay · yorky — Yorkie v1.1 Yorkshire dialect and accurate long-form writing. Plus myskills, which checks the five executable families and reports health and dependency readiness separately.
Paste this into your coding agent:
Install MOMM for me by following https://marroccofella.github.io/skills/momm/install
The signed install guide verifies the release, previews the exact skill and harness scope, and asks before installing or accepting protocol changes. It does not install the whole collection by default.
For contributors only: the default branch can contain unreleased work and this route does not verify a signed release. Inspect and trust its code separately before running even the preview. Do not use it as the stable MOMM installation path.
git clone https://github.com/marroccofella/skills && cd skills && node install.mjs --target claude --dry-run
The preview covers every skill for the named harness. Applying requires a separate decision; use the signed guide above for ordinary installation.
Invoke $mytravel to compare dates, airports, mixed cabins, stopovers, loyalty value and ground links against your exact requirements. Use available tools and authorised sessions, challenge the winner, and verify the booking and usable benefits.
Portable instructions with synthetic decision checks. Live prices, inventory and account access depend on your available tools. Scope and validation details →
Have other installed AI CLIs challenge your agent’s changes. MOMM 1.17.3 includes explicit signed updates, a separate protocol-acceptance gate and clearer setup, recovery and evidence pages. Your current agent remains the sole writer and verifier. Version history · Install or upgrade prompt
A historical snapshot of this project's development reviews: recorded runs, bounded normalized reviewer reports and governor decisions. Public hashes verify the sanitized bytes, not the truth of reviewer claims or the completeness of the project's history. No other user's private ledger is included.
Every momm user gets their own dashboard, generated locally from their own telemetry — it never touches this site or any server:
node momm/scripts/ledger.mjs --open
Separate per workspace and gitignored by protocol. It remains a local file, accessible to anyone with permission to read that folder; gitignore is not access control. Publishing is a separate explicit action. See the data boundaries.
Reviewer names identify harness CLIs authenticated by OAuth; inner model identities are not recorded. No cookies, no trackers. Part of the 42.uk universe · local voice workflows via Promptus. RELAX. IT'S ALREADY OVER.
Optional F5 narration: Built with Promptus · promptus.ai ↗ · A project by Professor Dominic Marrocco.