myskills
Skill0 starsChanged 40 days agoMITView on GitHub
Run the five canonical skill families from any harness and report functional health separately from dependency readiness — momm (multi-model peer review), myrepo (GitHub publishing), yorkshire-pudding/yorky (Yorkshire dialect), myautoness/autopilot (self-playing by search and verified replay), and promptus-clone-voice/myvoice (consented local voice cloning). Use when the user says "myskills", asks to check, verify, test, or list these skills, asks whether they are working, or wants the review → voice → publish flow. Reports aliases and versions without treating either as proof of health.
In this repository10 other files it gives its agents.
- dom-ttsSkill
- mommSkill
- multi-llm-reviewSkill
- myautonessSkill
- myrepoSkill
- mytravelSkill
- myvoiceSkill
- promptus-clone-voiceSkill
- yorkshire-puddingSkill
- yorkySkill
The file
Files
myskills — test the five skill families
One entry point that exercises each canonical family and reports both code health and the local dependencies needed to use it. A skill is reported working only when its own check passes — never because its folder exists or its version is current.
Run it
node scripts/run-all.mjs --pretty--pretty— human-readable JSON on stdout; a per-skill summary always prints to stderr.--quick— keep deterministic safety checks, skip live-dependency probes, and report skipped families or dependencies asnot_checked; it never turns unknown readiness green.--flow— print how the skills compose, without running anything.- Exit code is 1 when a functional check is missing, failing, or errored. Dependency-only attention, such as a missing engine or GitHub sign-in, is reported separately and does not falsely turn a passing safety test into a code failure.
Relay the per-skill lines to the user, including aliases, versions, dependency states, and any unavailable engine (for example, Promptus not installed). Never summarize code_health: passing as “everything is ready” when dependency_readiness still needs attention.
Canonical families and aliases
myskills 1.2.0 returns five functional records:
mommmyrepomyautoness, also invoked asautopilotyorkshire-pudding, also invoked asyorkypromptus-clone-voice, also invoked asmyvoice
An alias belongs on its canonical record; it is not another installed skill or another health check. myskills_version identifies the orchestrator itself and does not create a sixth functional family.
What it checks
| Skill | Called as | Proof required |
| momm | momm | every bundled deterministic --self-test check passes |
| myrepo | myrepo | zero-network --self-test privacy + secret gates pass; gh_cli and gh_auth readiness reported separately |
| myautoness | myautoness / autopilot | portable deterministic self-tests; gameplay replay requires a compatible reference adapter |
| yorkshire-pudding | yorky | self-tests pass and a live translation is verified |
| promptus-clone-voice | myvoice | Promptus services (ComfyUI, Cosy, CWorker) report ready |
Read the report
code_healthsummarizes functional checks. It isfailingonly when a canonical family's own proof is missing, failing, or errored.dependency_readinessisattentionwhen a required local engine, CLI, or account session is unavailable, andnot_checkedwhen--quickdeliberately skips those probes.- Each
skills[]record carries one canonicalskill, itsaliases, functionalstatus, explanatorydetail, and any detectedversion. - myrepo additionally reports
dependencies.gh_cliasreadyormissing, anddependencies.gh_authasready,login_required, ornot_checkedwhen the CLI is absent.
Version state and functional health answer different questions. A current version can still fail its self-test or lack a required engine; a local version can be newer than the published manifest and still pass. Do not group “current,” “update available,” or “local newer” as health verdicts, and do not relabel an unavailable version check as “up to date.”
How they compose
- momm — review it. Reproduce every finding before fixing; record dispositions.
- myvoice — voice it, with consent and a recorded human listening verdict.
- myrepo — publish it, after confirming repo name and visibility with the user.
- yorky — flavour any prose, never touching identifiers, keys, URLs, or logic.
Each skill keeps its own protocol and safety gates; myskills orchestrates and verifies, it never bypasses them. Install everything with node install.mjs --target all from the repo root.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

