nexus-tool-schemas
Skill158 starsChanged 11 days agoMITView on GitHub
Export the live CLI-first Nexus tool catalog as JSON — the whole registry or a selector-picked subset — and refresh the committed catalog that the repo's guidance and drift tests read. Use when someone asks for tool schemas or the exact command/argument shape a caller sees, after changing a tool's parameter schema, slug or description, after touching ToolCliNormalizer, or when a catalog-backed test fails.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- nexus-agentsSkill
- nexus-eval-harnessSkill
- nexus-llm-adaptersSkill
- nexus-mobile-compatSkill
- nexus-model-evalSkill
- nexus-model-updatesSkill
- nexus-releaseSkill
- nexus-storageSkill
- nexus-testingSkill
- nexus-ui-mockupsSkill
- nexusSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Run that protocol end to end.
- Distinctive: line 36
Run them, do not reimplement.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Nexus Tool Schemas
Context: the exporter boots the agent registry headlessly and asks the live
ToolCliNormalizer for each tool's CLI schema, so the JSON is the caller's-eye
view of the tools rather than a parse of the TypeScript. It runs with no
Obsidian, no vault and no build — but it does need npm install, and it writes
to two different files for two different jobs.
Workflow
- Decide which job this is. The answer is the output path, and choosing it by
default is the standard failure here — the default path is a scratch file
nothing reads.
- Reading, answering a question, inspecting a subset →
protocols/export-subset.md - Refreshing the catalog the repo ships and tests against →
protocols/refresh-catalog.md
- Reading, answering a question, inspecting a subset →
- Run that protocol end to end. You MUST NOT hand-write or hand-patch either JSON file: both are generated, and an edited catalog makes the drift test the only thing standing between a wrong doc and a caller.
- Validate what you produced with
scripts/check_catalog.pybefore reporting it, and treat a non-zero exit as a stop. - At the end of a session that used this skill, run
protocols/self-refine.md.
Map
protocols/the two export jobs, plus self-refine.references/consumers.md(which file each reader expects, and what breaks when it is stale) andexporter-internals.md(how the headless boot works, what it needs, and its failure modes).scripts/check_catalog.pyvalidates an exported catalog;scripts/check_exporter_coverage.pycatches an agent the exporter cannot see. Run them, do not reimplement.refinement-log.mdwhat past sessions changed here and why.
Siblings
- Adding or renaming an agent or tool, or a command that does not resolve:
nexus-agents. It owns the slug-to-CLI-name transform (cli_name.py) and the checker for commands written in docs (check_documented_commands.py). - Jest lanes, the shipped-docs gate and how to make a failing test fail for the
right reason:
nexus-testing. - Shipping the refreshed catalog in a version:
nexus-release.
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.

