Source profileQuality 74/100

LazyAGI/LazyMind/algorithm/lazymind/skills/create-plugin/SKILL.md

create-plugin

Use this skill **only** when the user explicitly asks to create a **new** plugin.

Source repository stars
55
Declared platforms
0
Static risk flags
0
Last source update
2026-07-28
Source checked
2026-07-28

Decision brief

What it does—and where it fits

Use this skill **only** when the user explicitly asks to create a **new** plugin.

Best for

  • "帮我创建一个…插件"
  • "新建一个插件"
  • "我想做一个插件"

Not for

  • Tasks that require unconfirmed production actions or broad system permissions.
  • Environments where the pinned source and install steps cannot be inspected.

Compatibility matrix

Platform support, with evidence labels

PlatformStatusEvidenceWhat to check
CodexNot declaredNo explicit evidencePortability before use
Claude CodeNot declaredNo explicit evidencePortability before use
CursorNot declaredNo explicit evidencePortability before use
Gemini CLINot declaredNo explicit evidencePortability before use
Open the compatibility checker

Installation

Inspect first. Install second.

The source command is displayed only when detected. A safe inspection prompt is always available so your agent can explain every action before execution.

Source-detected install commandSource
npx skills add https://github.com/LazyAGI/LazyMind --skill "algorithm/lazymind/skills/create-plugin"
Safe inspection promptEditorial

Inspect the Agent Skill "create-plugin" from https://github.com/LazyAGI/LazyMind/blob/b63cc44f8c68603bc9e802b56d8e243fef2481aa/algorithm/lazymind/skills/create-plugin/SKILL.md at commit b63cc44f8c68603bc9e802b56d8e243fef2481aa. List every install step, command, network request, credential, file read/write, external action, and rollback step. Explain whether it fits my task. Do not install or execute anything until I approve.

Workflow

What the source asks the agent to do

  1. 01

    WORKFLOW

    This skill follows a strict three-step flow. Do not skip steps or merge them.

    Slot type must be one of: text, image, file, json.Slot cardinality must be single or list.Keep steps to 2–5. Each step id should be a snakecase verb phrase (e.g. extractclauses).
  2. 02

    Step 1 — Generate a plugin summary (no tool call)

    Read the user's request carefully. Output a natural-language summary in this exact format:

    Slot type must be one of: text, image, file, json.Slot cardinality must be single or list.Keep steps to 2–5. Each step id should be a snakecase verb phrase (e.g. extractclauses).
  3. 03

    Step 2 — Handle user response

    If the user clicks Yes / 是 / 确认 (or any affirmative): proceed immediately to Step 3.

    If the user clicks Yes / 是 / 确认 (or any affirmative): proceed immediately to Step 3.If the user clicks No / 否 / 修改 or provides corrections: update the summary incorporating the feedback, then call askuser again. Repeat until the user confirms.- If the user clicks Yes / 是 / 确认 (or any affirmative): proceed immediately to Step 3. - If the user clicks No / 否 / 修改 or provides corrections: update the summary incorporating the feedback, then call askuser again. Re…
  4. 04

    Step 3 — Create the plugin draft (call createplugindraft)

    Call createplugindraft with: - name: the plugin display name from the confirmed summary. - description: a comprehensive description combining the functional description, slot details, and step list from the confirmed summary. Write enough detail for the AI to generate without am…

    name: the plugin display name from the confirmed summary.description: a comprehensive description combining the functional description, slot details, and step list from the confirmed summary. Write enough detail for the AI to generate without ambiguity.slots: the slot lines from the summary, one per line.

Permission review

Static risk signals and limitations

No configured static risk pattern was detected

This is not proof of safety. Runtime behavior, indirect dependencies, and hidden external systems are outside the static scan.

Evidence record

Why each signal appears

EvidenceSourceComputedTestedEditorial
SignalValueEvidence typeMeaning
Quality score74/100ComputedDocumentation, specificity, maintenance, and trust rules
Repository stars55SourceRepository attention, not individual Skill quality
Compatibility0 platformsSourceDeclared in the catalog source record
Usage guideautomated source guideEditorialGenerated or reviewed according to the visible evidence level

Pinned source

Provenance and original SKILL.md

Repository
LazyAGI/LazyMind
Skill path
algorithm/lazymind/skills/create-plugin/SKILL.md
Commit
b63cc44f8c68603bc9e802b56d8e243fef2481aa
License
Apache-2.0
Collected
2026-07-28
Default branch
main
View the original SKILL.md

Create Plugin Skill

WHEN TO USE

Use this skill only when the user explicitly asks to create a new plugin.

Trigger conditions:

  • "帮我创建一个…插件"
  • "新建一个插件"
  • "我想做一个插件"
  • Any similar request that clearly asks to author a brand-new plugin from scratch.

Do NOT use this skill when:

  • The user wants to modify or fix an existing plugin.
  • The user is asking about how plugins work (answer directly).
  • The user mentions a plugin in passing but is not requesting creation.

WORKFLOW

This skill follows a strict three-step flow. Do not skip steps or merge them.

Step 1 — Generate a plugin summary (no tool call)

Read the user's request carefully. Output a natural-language summary in this exact format:

插件名称:<display name>
功能描述:<one or two sentences describing what the plugin does>
输入槽位:<slot_id>(<type>,<cardinality>)<label>  — one per line
输出槽位:<slot_id>(<type>,<cardinality>)<label>  — one per line
主要步骤:1. <step_id>  2. <step_id>  3. <step_id>  ...

Rules for the summary:

  • Slot type must be one of: text, image, file, json.
  • Slot cardinality must be single or list.
  • Keep steps to 2–5. Each step id should be a snake_case verb phrase (e.g. extract_clauses).
  • Be specific enough that the AI generator can produce a working plugin without guessing.

After printing the summary, immediately call ask_user with:

{
  "questions": [
    {
      "id": "confirm",
      "type": "boolean",
      "text": "以上是插件方案,是否确认创建?如需调整请点击「修改」并说明修改意见"
    }
  ]
}

This suspends the turn and waits for user input.


Step 2 — Handle user response

  • If the user clicks Yes / 是 / 确认 (or any affirmative): proceed immediately to Step 3.
  • If the user clicks No / 否 / 修改 or provides corrections: update the summary incorporating the feedback, then call ask_user again. Repeat until the user confirms.

Step 3 — Create the plugin draft (call create_plugin_draft)

Call create_plugin_draft with:

  • name: the plugin display name from the confirmed summary.
  • description: a comprehensive description combining the functional description, slot details, and step list from the confirmed summary. Write enough detail for the AI to generate without ambiguity.
  • slots: the slot lines from the summary, one per line.
  • steps: the step ids from the summary, one per line.

After the tool returns successfully, write a brief confirmation message that:

  1. States the plugin is being generated.
  2. Includes a Markdown link to the editor using editor_url from the tool result: [点击这里打开插件编辑器](<editor_url>)
  3. Notes that generation runs in the background and the user can refresh to see progress.

Do NOT output any additional technical details or YAML.


IMPORTANT CONSTRAINTS

  • Never call create_plugin_draft before the user confirms (Step 2 must complete first).
  • Never call create_plugin_draft more than once per conversation request.
  • The ask_user tool is a stop-tool — after calling it the current turn ends. The user's next message is their answer; resume from Step 2.
  • Keep the summary in Chinese for readability, but slot ids and step ids must be English snake_case.

Alternatives

Compare before choosing