Ase Meta Quorum

---
name: ase-meta-quorum
argument-hint: "[--help|-h] [--models|-m <model>[,...]] <question>"
description: >
    Query multiple foreign AIs in parallel and synthesize a consensus
    ("quorum") answer with a consensus rating. Use when the user wants
    a quorum, a consensus answer, or the opinions of multiple AIs on a
    question.
user-invocable: true
disable-model-invocation: false
effort: high
allowed-tools:
    - "Agent"
    - "TaskCreate"
    - "TaskUpdate"
---

@${CLAUDE_SKILL_DIR}/../../meta/ase-control.md
@${CLAUDE_SKILL_DIR}/../../meta/ase-skill.md
@${CLAUDE_SKILL_DIR}/../../meta/ase-getopt.md

<purpose name="ase-meta-quorum">
Query Multiple AIs for Quorum Answer
</purpose>

<expand name="getopt"
    arg1="ase-meta-quorum"
    arg2="--models|-m=(all|chatgpt|gemini|deepseek|grok|glm|qwen|claude|codex|copilot)...">
    $ARGUMENTS
</expand>

<objective>
Find a *quorum answer* on an arbitrary question,
by querying *multiple* AIs for an *optimal consensus*.
</objective>

<flow>

1.  <step id="STEP 1: Preview Own Answer">

    The user-selectable foreign models are restricted by the
    `--models`/`-m` option, parsed into <getopt-option-models/>
    as a comma-separated list of model tokens. The getopt parser
    validates only the *first* token, so you *MUST* validate each
    remaining token yourself against the allowed set `all`, `chatgpt`,
    `gemini`, `deepseek`, `grok`, `glm`, `qwen`, `claude`, `codex`,
    `copilot`. If any token is *not*
    in this set, bind <token/> to that offending token, then only output
    the following <template/> and then immediately *STOP* processing the
    entire current skill:

    <template>
    ⧉ **ASE**: ✪ skill: **ase-meta-quorum**, ▶ ERROR: invalid `--models` token: **<token/>**
    </template>

    Prepare the LLM query by setting <query/> to the following <template/>:

    <template>
    <getopt-arguments/>.
    Please respond very concisely and briefly, with facts only,
    usually with just 1 to 7 corresponding bullet points and with short sentences.
    Optionally, mention potential cruxes which should be noticed.
    Besides bullet points, do not provide any additional explanations.
    Emphasize keywords or cruxes in your response with Markdown formatting.
    Format code parts with Markdown formatting.
    </template>

    For yourself (Anthropic Claude), first answer this <query/> in
    advance yourself by showing your own answer to the query as a sneak
    preview. For this, output the following <template/>:

    <template>
    **Anthropic Claude** (sneak preview in advance):
    - [...]
    - [...]
    </template>

    </step>

2.  <step id="STEP 2: Query Foreign AIs">

    The default is the single token `all`. If <getopt-option-models/>
    contains the token `all`, you *MUST* treat it as the full list
    `chatgpt,gemini,deepseek,grok,glm,qwen` (all API-based models). The
    harness-based models `claude`, `codex`, and `copilot` are *never*
    part of `all` and are queried only when their token is given
    *explicitly* (in addition to `all`, if wished). Anthropic
    Claude (yourself) is *always* included, independent of this option.

    <define name="agent">
    Call the `Agent` tool:

    ```text
        Agent(
            description:       "Query Foreign LLM: <arg1/>",
            subagent_type:     "ase:ase-meta-chat",
            prompt:            "<arg2/> <query/>",
            run_in_background: false
        )
    ```

    </define>

    Query only those foreign models whose token is contained in
    <getopt-option-models/> (where `all` selects every model); silently
    skip all others. Emit *all* these `Agent` tool invocations *in one
    single message* so they run in *parallel*:

    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `chatgpt`">
    <expand name="agent" arg1="OpenAI ChatGPT" arg2="chatgpt"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `gemini`">
    <expand name="agent" arg1="Google Gemini"  arg2="gemini"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `deepseek`">
    <expand name="agent" arg1="DeepSeek"       arg2="deepseek"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `grok`">
    <expand name="agent" arg1="xAI Grok"       arg2="grok"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `glm`">
    <expand name="agent" arg1="Z.AI GLM"       arg2="glm"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `all` OR <getopt-option-models/> contains `qwen`">
    <expand name="agent" arg1="Alibaba Qwen"   arg2="qwen"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `claude`">
    <expand name="agent" arg1="Anthropic Claude (harness)" arg2="claude"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `codex`">
    <expand name="agent" arg1="OpenAI Codex"               arg2="codex"></expand>
    </if>
    <if condition="<getopt-option-models/> contains `copilot`">
    <expand name="agent" arg1="GitHub Copilot"             arg2="copilot"></expand>
    </if>

    You *MUST* *NOT* output anything in this step.

    </step>

3.  <step id="STEP 3: Summarize Responses">

    Agents which returned a response with an `ERROR:` prefix are
    silently skipped and are treated as not available.

    Summarize all responses, of both yourself and all available agents,
    with just 1 to 7 corresponding bullet points and with short
    sentences.

    You *MUST* *NOT* output anything in this step.

    </step>

4.  <step id="STEP 4: Determine Consensus Rating">

    First, count the number of *available foreign AIs* (those queried in
    STEP 2 which did *not* return an `ERROR:` response). If this count is
    *zero* -- i.e., every foreign AI errored out or none were reachable
    and only yourself (Anthropic Claude) remains -- a quorum is *not*
    possible. In this case, set <c></c> and <n></n> to empty and instead
    set <disagreement/> to `(no quorum: no foreign AIs were available)`,
    then skip the rest of this step.

    Otherwise, let <n/> be the *total number of responders* (yourself plus
    all available foreign AIs above). Then determine, on a Likert scale of
    0..<n/>, the amount of the overall consensus <c/> of all the responses.
    If all responses disagree, the consensus <c/> is zero.
    If all responses agree, <c/> is <n/>.

    If not all AIs agree, determine <disagreement/> information,
    formatted as `(disagreement: <ai/>, <ai/>, [...])` where <ai/> is a
    name of an AI which disagreed with the consensus. Else, if all AIs
    agree, set <disagreement></disagreement>.

    You *MUST* *NOT* output anything in this step.

    </step>

5.  <step id="STEP 5: Show Results">

    Finally, show the summary, the consensus, and the complete and
    unmodified responses of yourself and each of the queried foreign
    AIs, based on the following output <template/>:

    <template>
    **QUESTION**:
    <getopt-arguments/>

    &#x25CF; **CONSENSUS ANSWER**:
    - [...]
    - [...]

    **CONSENSUS RATE**: **<c/>/<n/>** <disagreement/>

    &#x25CB; **Anthropic Claude**:
    - [...]
    - [...]

    &#x25CB; **OpenAI ChatGPT**:
    - [...]
    - [...]

    &#x25CB; **Google Gemini**:
    - [...]
    - [...]

    &#x25CB; **DeepSeek**:
    - [...]
    - [...]

    &#x25CB; **xAI Grok**:
    - [...]
    - [...]

    &#x25CB; **Z.AI GLM**:
    - [...]
    - [...]

    &#x25CB; **Alibaba Qwen**:
    - [...]
    - [...]

    &#x25CB; **Anthropic Claude (harness)**:
    - [...]
    - [...]

    &#x25CB; **OpenAI Codex**:
    - [...]
    - [...]

    &#x25CB; **GitHub Copilot**:
    - [...]
    - [...]
    </template>

    In this output, when a quorum was *not* possible (see STEP 4),
    render the consensus rate line instead as just `**CONSENSUS RATE**:
    *n/a* <disagreement/>` (omitting the `<c/>/<n/>` fraction), as a
    single AI cannot form a consensus.

    In this output, remove the sections of those AIs which were not
    queried (excluded via `--models`/`-m`) or were not available.
    You *MUST* *NOT* output any further explanations yourself.

    </step>

</flow>