A snapshot: read at commit 707a04c7ce8 on 2026-10-18T00:01:48Z by majordomus 0.29.0. Nothing on this page is read again. See it live
Skills
What an agent can be asked by name, and where each name comes from.
Workflow skills are written once under the layer; command and domain skills are derived from the capability registry. Every editor and agent reads the same projection.
- Workflow skills8skills.status /skills
- Projected files76editors.status /artifacts
- Editors and agents3editors.status /adapters
- Findings2editors.status /diagnostics
Workflow skills
The layer's own procedures, one directory each under the skills section.
| Skill | When to use it | Status |
|---|---|---|
| assess-before-deletingAssess before deleting | Remove a file, branch, worktree, record or process only after each candidate has been classified by what depends on it and what would be lost, and the reversible step has been taken first. | active |
| deploy-siteDeploy the site by hand | Publish the website to GitHub Pages from a terminal, through the one script the Pages workflow runs, when waiting for the Actions queue is not acceptable. | active |
| evolveEvolve the repository's know-how | Turn what sessions observed into reviewed knowledge and into the rule, skill, document or test each lesson belongs in, then deliver the change and tell every peer what moved. | active |
| implementImplement a change | Carry a requested change through the repository evidence-first, from the invariant it must keep to a verified, reconciled, documented and committed tree, without widening its scope. | active |
| pack-for-chatgptPack the sources for a ChatGPT project | Turn a committed tree into token-bounded Markdown shards that a ChatGPT project's file search indexes, verified to carry no binary, worktree, link, build output, projection or leaked value, and hand them over with the commit they were cut from. | active |
| pack-for-reviewPack the repository for an outside reviewer | Prepare a bounded, reproducible snapshot of the tracked tree for a reviewer who cannot read the repository directly, with the local half and anything secret left out and the questions the review must answer written down. | active |
| repo-reviewRepository review | An evidence-driven review of the actual state of the repository, or of one change to it, before meaningful engineering work is recommended or done. | active |
| report-verification-stateReport by verification state | Write the outcome of a piece of work as claims each marked VERIFIED, PARTIAL or BLOCKED by the evidence actually observed, so a reader can tell what was proved from what was only done or intended. | active |
skills.status387 msAPI/api/v1/skillsCLImajordomus skills statusMCPmajordomus_skillstechnical view
By name, in every editor and agent
Each name as it is typed, the adapter that reads it, and what it projects: a workflow skill, a capability, or a domain of the registry.
| Name | Read by | Projects | Support |
|---|---|---|---|
| AGENTS.md | Google Antigravity | rules | supported |
| /assess-before-deleting | Google Antigravity | skills | supported |
| /deploy-site | Google Antigravity | skills | supported |
| /evolve | Google Antigravity | skills | supported |
| /implement | Google Antigravity | skills | supported |
| /pack-for-chatgpt | Google Antigravity | skills | supported |
| /pack-for-review | Google Antigravity | skills | supported |
| /repo-review | Google Antigravity | skills | supported |
| /report-verification-state | Google Antigravity | skills | supported |
| /intent | Claude Code | commands | supported |
| /mj-artifacts | Claude Code | domains | supported |
| /mj-brief | Claude Code | domains | supported |
| /mj-capabilities | Claude Code | domains | supported |
| /mj-commands | Claude Code | domains | supported |
| /mj-commit | Claude Code | domains | supported |
| /mj-continuity | Claude Code | domains | supported |
| /mj-convergence | Claude Code | domains | supported |
| /mj-dashboard | Claude Code | domains | supported |
| /mj-delivery | Claude Code | domains | supported |
| /mj-deploy | Claude Code | domains | supported |
| /mj-design | Claude Code | domains | supported |
| /mj-devcontext | Claude Code | domains | supported |
| /mj-devtask | Claude Code | domains | supported |
| /mj-directories | Claude Code | domains | supported |
| /mj-distribution | Claude Code | domains | supported |
| /mj-economics | Claude Code | domains | supported |
| /mj-editors | Claude Code | domains | supported |
| /mj-entity | Claude Code | domains | supported |
| /mj-environment | Claude Code | domains | supported |
| /mj-episodes | Claude Code | domains | supported |
| /mj-evidence | Claude Code | domains | supported |
| /mj-executions | Claude Code | domains | supported |
| /mj-gates | Claude Code | domains | supported |
| /mj-graph | Claude Code | domains | supported |
| /mj-guard | Claude Code | domains | supported |
| /mj-health | Claude Code | domains | supported |
| /mj-integration | Claude Code | domains | supported |
| /mj-intent-realization | Claude Code | domains | supported |
| /mj-intents | Claude Code | domains | supported |
| /mj-knowledge-base | Claude Code | domains | supported |
| /mj-lifecycle | Claude Code | domains | supported |
| /mj-mcp | Claude Code | domains | supported |
| /mj-mesh | Claude Code | domains | supported |
| /mj-models | Claude Code | domains | supported |
| /mj-monitor | Claude Code | domains | supported |
| /mj-objects | Claude Code | domains | supported |
| /mj-obligations | Claude Code | domains | supported |
| /mj-observatory | Claude Code | domains | supported |
| /mj-pack | Claude Code | domains | supported |
| /mj-peers | Claude Code | domains | supported |
| /mj-perf | Claude Code | domains | supported |
| /mj-plan | Claude Code | domains | supported |
| /mj-product | Claude Code | domains | supported |
| /mj-quality | Claude Code | domains | supported |
| /mj-reasoning | Claude Code | domains | supported |
| /mj-recover | Claude Code | domains | supported |
| /mj-release | Claude Code | domains | supported |
| /mj-repository | Claude Code | domains | supported |
| /mj-rules | Claude Code | domains | supported |
| /mj-served | Claude Code | domains | supported |
| /mj-server | Claude Code | domains | supported |
| /mj-session-domain | Claude Code | domains | supported |
| /mj-shell | Claude Code | domains | supported |
| /mj-skills | Claude Code | domains | supported |
| /mj-trace | Claude Code | domains | supported |
| /mj-web | Claude Code | domains | supported |
| /mj-why | Claude Code | domains | supported |
| /mj-worktree | Claude Code | domains | supported |
| /assess-before-deleting | Claude Code | skills | supported |
| /deploy-site | Claude Code | skills | supported |
| /evolve | Claude Code | skills | supported |
| /implement | Claude Code | skills | supported |
| /pack-for-chatgpt | Claude Code | skills | supported |
| /pack-for-review | Claude Code | skills | supported |
| /repo-review | Claude Code | skills | supported |
| /report-verification-state | Claude Code | skills | supported |
| CLAUDE.md | Claude Code | instructions | supported |
| .mcp.json | Claude Code | mcp | supported |
| AGENTS.md | Zed | instructions | supported |
| /assess-before-deleting | Zed | skills | supported |
| /deploy-site | Zed | skills | supported |
| /evolve | Zed | skills | supported |
| /implement | Zed | skills | supported |
| /pack-for-chatgpt | Zed | skills | supported |
| /pack-for-review | Zed | skills | supported |
| /repo-review | Zed | skills | supported |
| /report-verification-state | Zed | skills | supported |
| .zed/tasks.json | Zed | tasks | supported |
editors.status196 msAPI/api/v1/editorsMCPmajordomus_editorstechnical view