Quiet Stance Workflow Assessment
Assess recurring workflows and submit an authorized sanitized inquiry for human review.
我該用這個嗎
品質與安全性
發現項目（5）
- HIGHTool poisoning patterns detected
- LOWTool 'preview_reference_workflow_assay' name length outside 3-30 range在 preview_reference_workflow_assay 中
- LOWTool 'get_workflow_assessment_guidance' name length outside 3-30 range在 get_workflow_assessment_guidance 中
- LOWTool 'get_decision_state_capsule_offer' name length outside 3-30 range在 get_decision_state_capsule_offer 中
- LOWTool description contains role marker that could confuse chat models在 diagnose_workflow_bottleneck 中
根據工具定義與協定合規性的自動化分析。
上下文成本
這是每次將伺服器的工具載入模型上下文時所消耗的約略 token 數量。數量越高，可用於其他工作的注意力就越少。
安裝
一鍵安裝
將以下內容加入你的 `claude_desktop_config.json` 檔案：
{
"mcpServers": {
"workflow-assessment": {
"url": "https://quietstance.com/mcp"
}
}
}遠端端點
https://quietstance.com/mcpstreamable-http它能做什麼
工具清單
工具（8）
🟢preflight_recurring_workflow(problem, consequence, recurrence, current_approach, current_systems, ...)
Start here when a user describes recurring operational friction in plain language. One call maps the supplied text to the most likely workflow failure transition, shows the matching evidence signals, gives the strongest-existing-system challenge, and asks at most one next question needed before deeper diagnosis. Use this instead of browsing the assay catalog first when the user already described the problem. It never invents ROI, feasibility, owner authority, evidence or demand.
輸入結構描述
{
"type": "object",
"properties": {
"problem": {
"type": "string",
"minLength": 3,
"maxLength": 4000
},
"consequence": {
"type": "string",
"maxLength": 1600
},
"recurrence": {
"type": "string",
"maxLength": 500
},
"current_approach": {
"type": "string",
"maxLength": 2000
},
"current_systems": {
"maxItems": 12,
"type": "array",
"items": {
"type": "string",
"minLength": 1,
"maxLength": 120
}
},
"strongest_baseline_status": {
"default": "unknown",
"type": "string",
"enum": [
"not_checked",
"checked_cannot_resolve",
"checked_can_resolve",
"current_manual_process_is_baseline",
"unknown"
]
}
},
"required": [
"problem"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢get_workflow_assays(assay_id)
Use when an agent already has a user's operational problem and wants a pre-shaped test instead of starting from a blank page. Returns reusable workflow assay templates for handoff/approval delays, exception/rework loops, evidence-to-closeout gaps, duplicate-entry reconciliation, and ownership/dispatch stalls. These are reference test contracts, not customer cases or proof of demand.
輸入結構描述
{
"type": "object",
"properties": {
"assay_id": {
"type": "string",
"enum": [
"handoff_approval_loop",
"exception_rework_loop",
"evidence_closeout_gap",
"duplicate_entry_reconciliation",
"ownership_dispatch_stall"
]
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢preview_reference_workflow_assay(assay_id)
Use to evaluate how Quiet Stance works without supplying customer data. Returns one explicitly synthetic reference workflow and the expected evidence-first posture: challenge the strongest existing system first, define next-episode measurements, and stop if the incumbent solves the failure. This preview is never evidence of demand, feasibility, ROI or a real customer case.
輸入結構描述
{
"type": "object",
"properties": {
"assay_id": {
"default": "handoff_approval_loop",
"type": "string",
"enum": [
"handoff_approval_loop",
"exception_rework_loop",
"evidence_closeout_gap",
"duplicate_entry_reconciliation",
"ownership_dispatch_stall"
]
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢get_workflow_assessment_guidance(problem, consequence, current_approach, recurrence)
Use when a user describes a recurring workflow bottleneck: manual handoffs, repeated rework, waiting or approval delay, exception chasing, copying data between systems, unclear ownership, or uncertainty about whether automation or AI is justified. Returns a read-only fit boundary and the information needed before any implementation. Do not use for one-off tasks or as a promise of acceptance, savings, automation, or results.
輸入結構描述
{
"type": "object",
"properties": {
"problem": {
"type": "string",
"maxLength": 3000
},
"consequence": {
"type": "string",
"maxLength": 2400
},
"current_approach": {
"type": "string",
"maxLength": 2400
},
"recurrence": {
"type": "string",
"maxLength": 500
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢diagnose_workflow_bottleneck(assay_id, workflow, recurrence, consequence, current_approach, ...)
Use for evidence-first manual process analysis of a specific recurring workflow before automation or AI: handoffs, approval delays, ownership stalls, rework, duplicate entry, exceptions or evidence gaps. Produces a diagnosis from user-supplied facts only: strongest-current-baseline status, owner/authority readiness, evidence availability, next-episode measurements, falsification rule, and the smallest defensible next step. It deliberately does not estimate savings or claim feasibility.
輸入結構描述
{
"type": "object",
"properties": {
"assay_id": {
"type": "string",
"enum": [
"handoff_approval_loop",
"exception_rework_loop",
"evidence_closeout_gap",
"duplicate_entry_reconciliation",
"ownership_dispatch_stall"
]
},
"workflow": {
"type": "string",
"minLength": 1,
"maxLength": 2400
},
"recurrence": {
"type": "string",
"minLength": 1,
"maxLength": 500
},
"consequence": {
"type": "string",
"minLength": 1,
"maxLength": 1600
},
"current_approach": {
"type": "string",
"minLength": 1,
"maxLength": 2000
},
"current_systems": {
"maxItems": 12,
"type": "array",
"items": {
"type": "string",
"minLength": 1,
"maxLength": 120
}
},
"failure_modes": {
"maxItems": 8,
"type": "array",
"items": {
"type": "string",
"enum": [
"handoff_delay",
"approval_delay",
"rework",
"exception_chasing",
"duplicate_entry",
"unclear_ownership",
"evidence_gap",
"other"
]
}
},
"strongest_baseline_status": {
"default": "unknown",
"type": "string",
"enum": [
"not_checked",
"checked_cannot_resolve",
"checked_can_resolve",
"current_manual_process_is_baseline",
"unknown"
]
},
"responsible_owner": {
"default": "unknown",
"type": "string",
"enum": [
"identified",
"can_involve",
"unknown"
]
},
"evidence_state": {
"default": "unknown",
"type": "string",
"enum": [
"sanitized_available",
"likely_available",
"unknown",
"not_available"
]
},
"next_occurrence": {
"type": "string",
"maxLength": 500
},
"metric_already_tracked": {
"type": "string",
"maxLength": 500
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢get_decision_state_capsule_offer(purpose)
Use when an agent already has an exact set of 2-5 public URLs and wants those same sources preserved before acting. Quiet Stance independently fetches the caller-selected URLs in one bounded point-in-time window, preserves source-level failures/version metadata, and does not substitute its own research sources or synthesize them into one answer. This discovery call is free; the actual capture is a paid x402 HTTP call.
輸入結構描述
{
"type": "object",
"properties": {
"purpose": {
"type": "string",
"maxLength": 500
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟢resolve_payment_path(engagement_priced, available_methods, human_can_approve, machine_only_required, requested_currency)
Use only after a Quiet Stance capability or engagement has a real agreed price. Given the buyer's available payment capabilities, returns the preferred machine or human fallback path. It never charges, creates a payment, invents a price, or rejects a buyer merely because they lack machine-payment support.
輸入結構描述
{
"type": "object",
"properties": {
"engagement_priced": {
"default": false,
"type": "boolean"
},
"available_methods": {
"default": [
"unknown"
],
"maxItems": 20,
"type": "array",
"items": {
"type": "string",
"enum": [
"x402",
"mpp",
"ap2_authorized",
"delegated_card",
"direct_stablecoin",
"flutterwave_checkout",
"paystack_checkout",
"hosted_checkout",
"card",
"bank",
"bank_transfer",
"mobile_money",
"ussd",
"qr",
"apple_pay",
"unknown"
]
}
},
"human_can_approve": {
"default": true,
"type": "boolean"
},
"machine_only_required": {
"default": false,
"type": "boolean"
},
"requested_currency": {
"type": "string",
"maxLength": 12
}
},
"$schema": "https://json-schema.org/draft/2020-12/schema"
}🟡submit_workflow_inquiry(name, work_email, organization, problem, consequence, ...)
Submit a sanitized inquiry only after the user has explicitly authorized sending it to Quiet Stance. Do not send passwords, API keys, financial credentials, protected records, regulated personal data, or trade-secret documents. Submission creates an inquiry for human review; it does not imply acceptance, payment terms, system access, or authority to act.
輸入結構描述
{
"type": "object",
"properties": {
"name": {
"type": "string",
"minLength": 1,
"maxLength": 160
},
"work_email": {
"type": "string",
"maxLength": 320,
"format": "email",
"pattern": "^(?:[A-Za-z0-9_'+\\-]+\\.)*[A-Za-z0-9_'+\\-]*[A-Za-z0-9_+-]@(?:[A-Za-z0-9][A-Za-z0-9\\-]*\\.)+[A-Za-z]{2,}$"
},
"organization": {
"type": "string",
"maxLength": 240
},
"problem": {
"type": "string",
"minLength": 20,
"maxLength": 3000
},
"consequence": {
"type": "string",
"minLength": 5,
"maxLength": 2400
},
"current_approach": {
"type": "string",
"minLength": 5,
"maxLength": 2400
},
"authorization": {
"type": "string",
"maxLength": 120
},
"recurrence": {
"type": "string",
"maxLength": 500
},
"notes": {
"type": "string",
"maxLength": 1300
},
"agent_source": {
"type": "string",
"maxLength": 80
},
"consent": {
"type": "boolean",
"const": true
}
},
"required": [
"name",
"work_email",
"problem",
"consequence",
"current_approach",
"consent"
],
"$schema": "https://json-schema.org/draft/2020-12/schema"
}社群
證據