Octo Marketplace logo

Octo Marketplace

Organization
Mininglamp-OSS
octo-marketplace

Search, install, publish, review, and update Marketplace plugins — Skills, MCP connectors, and Experts/Squads (专家/专家团) — through the unified plugin API. Load after octo-shared.

Overview

PublisherMininglamp-OSS
Repositoryocto-cli
Skill nameocto-marketplace
Stars
918
Forks
130
Bundled files
3
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Mininglamp-OSS on GitHub. Read the source before you install it.

Installation

Install the Octo Marketplace AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Mininglamp-OSS/octo-cli/tree/main/skills/octo-marketplace
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Mininglamp-OSS/octo-cli.git /tmp/octo-cli
mkdir -p .claude/skills
cp -r /tmp/octo-cli/skills/octo-marketplace .claude/skills/octo-marketplace
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Octo Marketplace in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Octo Marketplace on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Octo Marketplace is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

octo-marketplace — unified plugin catalog

The marketplace is one backend with a single catalog surface: every asset is a plugin with a plugin_type. All commands authenticate with the active credential and send its Space context; a bf_* User Bot is resolved to its owner and authoritative Bot Space by the backend. Unified catalog operations use /market/api/v1/plugins*; helper operations such as categories, uploads, probes, and icon presigns use other routes under /market/api/v1/*.

plugin_typewasread via
skillSkill ZIPskills.md
connectorMCP listingmcp.md
expertExpert (专家)expert.md
expert_teamSquad (专家团)expert.md
bash
octo-cli marketplace plugin list --scene-code default --plugin-type skill --q "<kw>"
octo-cli marketplace plugin get --plugin-id <id> --include-relations

scene-code is default. Catalog listing requires --scene-code and --plugin-type; the owned all-types view is the exception:

bash
octo-cli marketplace plugin list --scene-code default --mode mine

Use immutable plugin_id and review_id values, never names.

Save, publish, and review lifecycle

Creating with plugin upsert or skill plugin import saves a draft and a version snapshot; it does not expose the plugin to the Space catalog. Updates preserve the current listing state: draft/delisted content remains editable, published private content remains owner-only, and published Space content rejects direct edits and must use review. Version labels use MAJOR.MINOR.PATCH (each numeric component is 1–9 digits) and may not move backward.

After verifying a saved draft, publish it explicitly:

bash
octo-cli marketplace plugin publish --plugin-id <plugin-id> \
  --version 1.2.0 --changelog "What changed"
  • A private plugin's listing state becomes published immediately, but it stays owner-only and is not exposed in the Space catalog.
  • A space plugin opens a review request and remains a draft until a Space owner/admin approves it. Save the returned review_id.
  • Read display_status as the user-facing state: draft, pending_review, published, rejected, or delisted. listing_state is the lower-level listing axis.

Applicants can inspect or cancel their requests:

bash
octo-cli marketplace plugin review-request list --mode mine --status pending
octo-cli marketplace plugin review-request get <review-id>
octo-cli marketplace plugin review-request cancel <review-id>

Space owners/admins review the frozen submission, not mutable live draft data:

bash
octo-cli marketplace plugin review-request list --mode space --status pending
octo-cli marketplace plugin review-request get <review-id>
octo-cli marketplace plugin review-request approve <review-id>
octo-cli marketplace plugin review-request reject <review-id> --reason "Explain the required correction"

A Space owner/admin can take published Space content down; this leaves the plugin editable and eligible for a later publish:

bash
octo-cli marketplace plugin delist --plugin-id <plugin-id> --reason "Policy or maintenance reason"

An already-published Space plugin cannot be edited through upsert or skill import; the backend rejects that bypass. Submit the full frozen upgrade with review-request create: a listed upgrade must carry either parse_task_id for a freshly parsed skill archive or both manifest_json + plugin_json for declared documents. Only an unlisted Space-intent first submission may omit content and snapshot its live draft. A present relations array replaces the frozen graph; omitting it inherits the live graph.

Retry and recovery

Marketplace mutations are not transport-retried. A gateway failure may mean the server committed the operation, so RESULT_UNKNOWN must be resolved by reading state before retrying:

  • create: for Skills, walk plugin list --scene-code default --plugin-type skill --mode mine --page 1 --page-size 100 through the complete owned Skill list, re-passing --page-size 100 on every page, and compare manifest_json.name against the stable machine name. Follow the detailed procedure in skills.md. For experts, expert teams, and connectors, walk plugin list --scene-code default --plugin-type <type> --mode mine --page 1 --page-size 100 until a short page is returned, re-passing --page-size 100 on every page, and compare each row's plugin_name exactly with the title submitted by the create request. Do not assume manifest_json.name is a machine identity for those non-Skill types. For every asset type, if multiple rows match exactly, stop as ambiguous and ask the user which plugin_id is authoritative; never pick a row automatically. Retry the create only after the relevant scan reaches a short page with zero exact matches;
  • update/import of an existing id: plugin get --plugin-id <id> and compare the intended version, hashes, and content; use plugin version list where useful;
  • publish: plugin get --plugin-id <id> and inspect display_status / review_id, then list owned review requests;
  • review decisions: re-read the review request;
  • delist/delete: re-read the plugin.

--q matches human-facing plugin_name and manifest-description substrings, not the stable manifest_json.name; a page defaults to 20 rows. Keep the same explicit --page-size on every request, and continue through a short page even after finding an exact match. Only then decide whether the match is unique; never stop at the first exact match.

Pagination and filtering

plugin list uses --page / --page-size (default 20, max 100); there is no --page-all. Filters are --category-id, --q, repeatable --tag, --mode mine, and --sort (newest, oldest, updated, name, placement, views, installs, downloads, comprehensive).

plugin-category list requires --scene-code and --plugin-type. plugin-tag list accepts optional --scene-code, --plugin-type, --q, --mode mine, and --limit (default 50, max 100).

Load only the reference matching the task: connector mcp.json + probe, skill attachment tree + upload/import, or expert/team AGENTS.md + relations.

Shared output rule

For a non-paginated command, normalize before reading fields:

bash
payload=$(octo-cli marketplace plugin get --plugin-id <id> | jq -c '.data.data // .data')

Do not apply that to {data:[...], pagination:{...}}; the CLI exposes items at .data and pagination at ._pagination.

Shared safety rule

  • Show the target and intended mutation before create, update, publish, review decision, delist, install, or delete; continue only after explicit user confirmation.
  • Never print presigned URLs, tokens, secret-bearing headers, or raw secret values. Connector secrets use ${VAR} placeholders.
  • Do not modify a plugin or review request outside the user-selected target.
  • Preserve the prior runtime configuration if installation fails.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Octo Marketplace AI skill do?

Search, install, publish, review, and update Marketplace plugins — Skills, MCP connectors, and Experts/Squads (专家/专家团) — through the unified plugin API. Load after octo-shared.

Why use Octo Marketplace on TypingMind?

Because you install it once and use it with any model. Octo Marketplace is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Octo Marketplace in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Mininglamp-OSS/octo-cli/tree/main/skills/octo-marketplace. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Octo Marketplace?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Octo Marketplace?

As many as you like. As long as a model supports skills, you can use Octo Marketplace with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Octo Marketplace AI skill free?

Yes. It is published on GitHub by Mininglamp-OSS under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇