Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions plugins/Coding-jpg/glasser/.claude-plugin/plugin.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
{
"name": "glasser",
"version": "0.1.0",
"description": "Discover and inspect paid data API endpoints with the Glasser CLI, then execute only user-approved requests.",
"author": {
"name": "Xinwei Liao",
"url": "https://github.com/Coding-jpg"
},
"homepage": "https://glasser.ai/",
"repository": "https://github.com/glasser-ai/plugins",
"license": "MIT",
"keywords": [
"data research",
"seo",
"company research",
"glasser"
],
"skills": [
"./skills/SKILL.md"
]
}
21 changes: 21 additions & 0 deletions plugins/Coding-jpg/glasser/LICENSE
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
MIT License

Copyright (c) 2026 glasser.ai

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
40 changes: 40 additions & 0 deletions plugins/Coding-jpg/glasser/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
# Glasser research Skill for MiniMax Code

A CLI-oriented community adaptation of the official [Glasser plugin](https://github.com/glasser-ai/plugins), prepared by Xinwei Liao (Glasser GTM Team). This package is maintained under Coding-jpg; it is not a MiniMax endorsement. Upstream copyright and MIT license are retained in LICENSE.

## User problem

Research agents need to find a suitable paid data endpoint, understand its input contract and current cost, and keep execution within the user's approved scope. This Skill documents that workflow using the official Glasser CLI.

## Example prompt and expected result

> Find a suitable Glasser endpoint for competitor keyword research for the domain and market I provide. Show me its schema and current price. Do not execute any paid request until I approve it.

Expected result: candidate endpoints, one inspected contract, and a proposed small request with pricing. The Skill itself does not perform a paid call or fabricate research results.

## Installation and dependencies

For MiniMax Code 0.7.2, copy this directory to `<MiniMax data directory>/plugins/glasser`, then run `mcode plugin list --marketplace local --available --json` to verify discovery. The data directory can be set with `MINIMAX_DATA_DIR`; keep validation separate from an existing installation. The package includes the portable Agent Plugins 1.0 manifest and the MiniMax Code 0.4+ manifest, with byte-identical copies of one Skill. It has no hooks, custom agents, commands or bundled MCP servers.

The host needs a shell and an already installed official Glasser CLI (`@glasser-ai/cli`), Node.js 22+, internet access, and an authenticated Glasser account/Workspace. Refer to the [official upstream setup instructions](https://github.com/glasser-ai/plugins) for CLI installation and login. The package does not install executables. Host/client and CLI platform support apply; no claim of testing every OS is made.

## Accounts and paid services

A Glasser account and Workspace are required for authenticated use. Data requests are pay-as-you-go. Inspect the endpoint's current price and charge clauses before requesting user approval. The MIT license applies to these instructions, not to hosted data services. MiniMax account or model charges are separate.

## Network and data behavior

This package makes no network requests on load. When the user invokes the installed CLI, it contacts Glasser's documented API and authentication endpoints, including `https://api.glasser.ai` and `https://app.glasser.ai`. Requests may be routed to the selected third-party data provider. The host client may process prompts and returned results. No credentials, private endpoints, telemetry, installers, binaries or symlinks are included.

## Documentation, privacy and support

- Product documentation: https://glasser.ai/docs
- Official source: https://github.com/glasser-ai/plugins
- Example workflow: https://glasser.ai/tutorial/analyze-competitor-keywords
- Support: support@glasser.ai
- Privacy policy: https://glasser.ai/privacy-policy
- Service terms: https://glasser.ai/terms-of-service

## Validation scope

Validated on macOS with MiniMax Code 0.7.2 using an isolated local data directory: the CLI reports `glasser@local`, enabled, with one Skill and no MCP server. The repository package validator passes. The full repository suite ran 1,121 tests: 1,118 passed, two unrelated existing-plugin tests failed in this macOS environment, and one was skipped (see the submission report). No model conversation or paid Glasser request was executed; those end-to-end results are not claimed.
19 changes: 19 additions & 0 deletions plugins/Coding-jpg/glasser/plugin.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,19 @@
{
"$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
"name": "glasser",
"version": "0.1.0",
"description": "Discover and inspect paid data API endpoints with the Glasser CLI, then execute only user-approved requests.",
"author": {
"name": "Xinwei Liao",
"url": "https://github.com/Coding-jpg"
},
"homepage": "https://glasser.ai/",
"repository": "https://github.com/glasser-ai/plugins",
"license": "MIT",
"keywords": [
"data research",
"seo",
"company research",
"glasser"
]
}
59 changes: 59 additions & 0 deletions plugins/Coding-jpg/glasser/skills/SKILL.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
---
name: glasser
description: Use for paid data research, company research, SEO, web search or local business data when the task requires a provider that is not already available. Use the official Glasser CLI to discover endpoints and inspect current prices before any approved paid request.
---

# Research with Glasser CLI

This is a CLI-oriented adaptation of the official Glasser plugin for MiniMax
Code. It contains instructions only. It does not bundle an MCP connection,
a server, an installer, or credentials. Official upstream:
https://github.com/glasser-ai/plugins .

## Prerequisites

Use an already installed official Glasser CLI. The official package is
`@glasser-ai/cli` and requires Node.js 22 or newer. Refer to the upstream
setup instructions if it is missing. Do not run installation commands
without the user's authorization. Follow the official documentation at
https://glasser.ai/docs and https://glasser.ai/docs/mcp-server for product
behavior. Do not assume MiniMax Code supports generic OAuth MCP setup.

## Workflow

1. Confirm the user's question, target and scope. If an existing tool already
provides the needed capability, use it first.
2. Check authentication with `glasser balance`. If the CLI reports missing
or rejected authentication, use the documented `glasser login` flow with
the user. Never request a key in chat or commit it to a file.
3. Discover endpoints: `glasser search -q "<research capability>"`.
4. Inspect a candidate: `glasser inspect -p <provider> -e <endpoint>`.
Read its inputs, current price and outcome-specific charge conditions.
5. Present the proposed request and price and obtain approval before any paid
execution. Use the CLI's own help and the inspected schema for the exact
run syntax; do not invent options. Start with small input scope.
6. Report the actual outcome, provider, charge and limitations. Follow the
documented run status handling for the selected endpoint. Do not perform
speculative paid retries, bulk runs or scope expansion.

## Example

Ask: "Find a suitable endpoint for competitor keyword research for the domain
and market I provide. Show the schema and price first. Do not execute a paid
request until I approve it."

Expected result before approval: endpoint candidates, an inspected input
contract and a proposed small request with current pricing. No paid data
request should run merely to demonstrate that the Skill is loaded.

## Data and billing

Authenticated data requests go to Glasser and its selected provider. The
MiniMax host may process prompts and returned data under its own policies.
Use only task-appropriate authorized data. The plugin is MIT-licensed;
Glasser's service is separately metered. Empty and failed outcomes follow
the selected endpoint's charge terms, not a blanket free-call promise.

Support: support@glasser.ai
Privacy: https://glasser.ai/privacy-policy
Service terms: https://glasser.ai/terms-of-service
59 changes: 59 additions & 0 deletions plugins/Coding-jpg/glasser/skills/glasser/SKILL.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,59 @@
---
name: glasser
description: Use for paid data research, company research, SEO, web search or local business data when the task requires a provider that is not already available. Use the official Glasser CLI to discover endpoints and inspect current prices before any approved paid request.
---

# Research with Glasser CLI

This is a CLI-oriented adaptation of the official Glasser plugin for MiniMax
Code. It contains instructions only. It does not bundle an MCP connection,
a server, an installer, or credentials. Official upstream:
https://github.com/glasser-ai/plugins .

## Prerequisites

Use an already installed official Glasser CLI. The official package is
`@glasser-ai/cli` and requires Node.js 22 or newer. Refer to the upstream
setup instructions if it is missing. Do not run installation commands
without the user's authorization. Follow the official documentation at
https://glasser.ai/docs and https://glasser.ai/docs/mcp-server for product
behavior. Do not assume MiniMax Code supports generic OAuth MCP setup.

## Workflow

1. Confirm the user's question, target and scope. If an existing tool already
provides the needed capability, use it first.
2. Check authentication with `glasser balance`. If the CLI reports missing
or rejected authentication, use the documented `glasser login` flow with
the user. Never request a key in chat or commit it to a file.
3. Discover endpoints: `glasser search -q "<research capability>"`.
4. Inspect a candidate: `glasser inspect -p <provider> -e <endpoint>`.
Read its inputs, current price and outcome-specific charge conditions.
5. Present the proposed request and price and obtain approval before any paid
execution. Use the CLI's own help and the inspected schema for the exact
run syntax; do not invent options. Start with small input scope.
6. Report the actual outcome, provider, charge and limitations. Follow the
documented run status handling for the selected endpoint. Do not perform
speculative paid retries, bulk runs or scope expansion.

## Example

Ask: "Find a suitable endpoint for competitor keyword research for the domain
and market I provide. Show the schema and price first. Do not execute a paid
request until I approve it."

Expected result before approval: endpoint candidates, an inspected input
contract and a proposed small request with current pricing. No paid data
request should run merely to demonstrate that the Skill is loaded.

## Data and billing

Authenticated data requests go to Glasser and its selected provider. The
MiniMax host may process prompts and returned data under its own policies.
Use only task-appropriate authorized data. The plugin is MIT-licensed;
Glasser's service is separately metered. Empty and failed outcomes follow
the selected endpoint's charge terms, not a blanket free-call promise.

Support: support@glasser.ai
Privacy: https://glasser.ai/privacy-policy
Service terms: https://glasser.ai/terms-of-service