hb-wrap
Skill2 starsChanged 22 days agoMITView on GitHub
Use Hypagon REST wrappers as deterministic, authenticated REST tools without rebuilding URLs or handling credentials.
In this repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- daisyuiSkill
- jpl-code-reviewerSkill
- hb-browserSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 2 are distinctive, and 4 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.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Use this skill when an MCP client has the hb-wrap server available.
- Distinctive: line 48
Call it with only the declared flat arguments:
- Very specific: line 54
If the wrapper is exposed as a generated tool, call wrap_orders_create with the same flat arguments.
- Very specific: line 63
Always run wrap_get before wrap_run.
- Very specific: line 89
Use the returned recipe as the response field of wrap_create or wrap_update, then run wrap_run against a real request.
- Distinctive: line 102
If the secret is missing or rejected, ask the human to refresh it.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Hypagon REST
Use this skill when an MCP client has the hb-wrap server available.
Install:
hermes skills install https://raw.githubusercontent.com/hyperbolictonic/hypagon/main/skills/hb-wrap/SKILL.md
npx skills add hyperbolictonic/hypagon --skill hb-wrapThe app must be running. The wrapper service talks to the local bridge at
http://127.0.0.1:4111 and requires the generated wrapper token.
Tool surface
wrap_list— list saved wrapperswrap_get— inspect one wrapper and its input schemawrap_run— execute a wrapper with flat semantic argumentswrap_test— dry-run the request or render a sample response; sends nothingwrap_derive— infer a first-draft response recipe from JSON or a URLwrap_create— save a wrapperwrap_update— replace a wrapperwrap_delete— delete a wrapperwrap_secret_list— list credential names; values are never availablewrap_reload— reload the manifest- generated
wrap_<id>tools — exposed wrappers with their own schemas
Recipes
Discover and call a wrapper
{"name":"wrap_list","arguments":{}}Inspect the chosen wrapper:
{"name":"wrap_get","arguments":{"id":"orders.create"}}Call it with only the declared flat arguments:
{"name":"wrap_run","arguments":{"id":"orders.create","args":{"customer_id":"cus_123","quantity":2}}}If the wrapper is exposed as a generated tool, call wrap_orders_create with
the same flat arguments.
wrap_run returns projected data, not a page. Context economy comes from the
response recipe (maxItems, maxChars, maxFieldChars) and the pagination
cap, not from JEV; JEV only applies to browser page reads.
Inspect before writing
Always run wrap_get before wrap_run. Do not guess argument names or assemble
a URL yourself. The wrapper owns the URL, method, query, headers and body
templates. Verify a new or edited wrapper with wrap_test (dry-run is the
default) before spending a live request.
Dry-run before the live call
Preview the exact request without sending it:
{"name":"wrap_test","arguments":{"id":"orders.create","args":{"customer_id":"cus_123","quantity":2}}}wrap_test defaults to dryRun: true. After editing a recipe, pass a sample
response to check the rendering instead of guessing:
{"name":"wrap_test","arguments":{"id":"orders.create","sample":{"items":[{"id":1,"name":"A"}]}}}Derive a recipe from a sample
{"name":"wrap_derive","arguments":{"sample":{"items":[{"id":1,"name":"A"}]},"name":"items"}}Use the returned recipe as the response field of wrap_create or
wrap_update, then run wrap_run against a real request.
Credentials
List credential names:
{"name":"wrap_secret_list","arguments":{}}A wrapper references a credential by name in request.auth. Never request or
print the credential value. A human must bind the value in Settings → Wrappers.
If the secret is missing or rejected, ask the human to refresh it.
Create a wrapper
wrap_create takes one complete object:
{
"wrapper": {
"schemaVersion": 1,
"id": "example.items",
"name": "Example items",
"description": "List items from the example API.",
"request": {
"method": "GET",
"url": "https://api.example.com/items",
"auth": "example-key"
},
"inputSchema": {
"type": "object",
"properties": {
"limit": { "type": "integer", "minimum": 1, "maximum": 100, "default": 20 }
},
"additionalProperties": false
},
"response": {
"schemaVersion": 1,
"select": "$.items[*]",
"mode": "table",
"fields": {
"Name": "name",
"ID": "id"
},
"record": { "id": "id", "name": "name" },
"limits": { "maxItems": 20 }
},
"exposeAsTool": false,
"mutating": false
}
}Then dry-run it with wrap_test, check the rendered sample, and only then rely
on wrap_run.
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.

