用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/UitbreidenOS/UitKit --skill process-mapper命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Guidelines and instructions for Agent execution state rollback rules
Guidelines and instructions for Agent execution step counters limits
Guidelines and instructions for Agent execution timeout limits setups
基于 SOC 职业分类
| name | process-mapper |
| description | Map and document business processes: flowchart, RACI, bottleneck identification, improvement recommendations |
/sop-writer to update itMap the following business process end-to-end.
Process name: [e.g., Customer onboarding, Invoice approval, Vendor procurement]
Trigger: [What starts this process? e.g., New customer signs contract]
End state: [What does "done" look like? e.g., Customer has logged in and completed setup]
Participants: [List all roles involved — e.g., Sales, Customer Success, Finance, IT]
Tools/systems involved: [CRM, ERP, email, Slack, etc.]
Known pain points (if any): [what you already know is broken or slow]
Produce:
## 1. Process overview
- Start trigger
- End state
- Estimated total duration (best case / worst case)
- Number of handoffs
- Systems touched
## 2. Step-by-step process map
For each step:
- Step name
- Who does it (role, not person)
- Input: what they receive
- Output: what they produce
- Tool/system used
- Estimated time
- Common failure mode
Format as a numbered table with columns: # | Step | Owner | Input | Output | Tool | Time | Failure Mode
## 3. RACI matrix
Map each step to: Responsible / Accountable / Consulted / Informed
Rules:
- Only ONE person can be Accountable per step (if multiple, that's a problem)
- Responsible does the work. Accountable owns the outcome. Don't confuse them.
- Consulted = must be asked before action. Informed = told after.
## 4. Bottleneck analysis
Identify steps where:
- Cycle time is disproportionate to value added
- Handoffs fail or are delayed most often
- The same rework or error occurs repeatedly
- A single person is a chokepoint (key-person dependency)
Score each step: Green (smooth) / Amber (delays common) / Red (frequent failures)
## 5. Improvement recommendations
For each Red/Amber step:
- Root cause (why does this fail?)
- Quick fix (< 1 week, no new tools)
- Medium fix (2-4 weeks, may require new tooling)
- Process owner for the fix
- Estimated impact: [time saved / error rate reduction / cost saved]
## 6. Automation opportunities
Which steps are candidates for automation?
Criteria: repetitive, rule-based, high volume, low judgment required
For each candidate: what tool/system would handle it, and what's the ROI estimate?
Give me a quick process map for: [PROCESS NAME]
Context:
- Who triggers it: [role]
- Who finishes it: [role]
- Approximate steps: [N]
- Main handoff points: [list them]
Output:
1. Linear step list with owner and tool for each step
2. Top 2 bottlenecks (where it usually breaks)
3. One improvement recommendation
Build a RACI matrix for the following process.
Process: [name]
Steps: [list each step, numbered]
Roles involved: [list all roles, e.g., Ops Manager, Finance, IT, Legal, CEO]
Rules to apply:
1. Exactly one Accountable per step — if you'd put two, flag it as an ownership problem
2. If Consulted > 3 per step, flag it as decision bottleneck
3. If a role appears as Accountable on > 50% of steps, flag key-person dependency
Output:
- Full RACI table (rows = steps, columns = roles)
- Ownership problems identified (shared accountability / no accountability)
- Bottleneck roles (overloaded with Responsible or Consulted)
- Recommendation: which role should own this process end-to-end?
I have a process with a known bottleneck at step: [STEP NAME]
What we know:
- Average time this step takes: [X hours/days]
- Expected time it should take: [X hours/days]
- Who owns it: [role]
- What causes delays (from observation): [list known causes]
- Downstream impact of delay: [what happens when this is late]
Run a 5 Whys analysis on this bottleneck:
Why 1: [why does it take longer than it should?]
Why 2: [why does that happen?]
...down to root cause
Then produce:
- Root cause statement (1 sentence)
- 3 intervention options (quick / medium / structural)
- Recommendation with rationale
- Success metric: how will we know it's fixed?
interface ProcessStep {
id: number
name: string
owner: string // role, not person
input: string
output: string
tool: string
estimatedMinutes: number
failureMode: string
bottleneckRating: 'green' | 'amber' | 'red'
}
interface RACIEntry {
stepId: number
stepName: string
roles: Record<string, 'R' | 'A' | 'C' | 'I' | '-'>
}
interface ProcessMap {
name: string
trigger: string
endState: string
totalDurationMin: number
totalDurationMax: number
handoffCount: number
steps: ProcessStep[]
raci: RACIEntry[]
: {
:
:
:
:
}[]
: {
:
:
:
}[]
}
Convert the following process steps into a Mermaid flowchart.
Steps:
1. [Step name] — Owner: [role]
2. [Step name] — Owner: [role]
...
Rules:
- Use different shape per role (rectangle = one department, diamond = decision/approval)
- Decision points (yes/no, approve/reject) must use diamond shapes
- Show handoffs as arrows with labels
- Mark bottleneck steps in red with a comment
Output the complete mermaid diagram code block.
Example output:
flowchart TD
A([Customer signs contract]) --> B[Sales: Update CRM & notify CS]
B --> C{CS: Onboarding ready?}
C -- No --> D[CS: Request missing info]
D --> C
C -- Yes --> E[CS: Send welcome email]
E --> F[IT: Provision accounts]
F:::bottleneck --> G[CS: Schedule kickoff call]
G --> H([Customer onboarded])
classDef bottleneck fill:#ff4444,color:#fff
User: Map our vendor procurement process. It starts when a team lead requests a new tool and ends when the vendor is contracted and access is provisioned. Involves: Team Lead, Ops Manager, Finance, Legal, IT.
Output:
## Process Overview
Start: Team Lead submits tool request
End: Vendor contracted, user access provisioned
Duration: Best case 5 days / Worst case 6 weeks
Handoffs: 7
Systems: Request form (Google Form), email, DocuSign, IT ticketing
## Step-by-Step Map
| # | Step | Owner | Input | Output | Tool | Time | Failure Mode |
|---|---|---|---|---|---|---|---|
| 1 | Submit tool request | Team Lead | Business need | Completed form | Google Form | 30 min | Vague justification, no ROI |
| 2 | Ops review & triage | Ops Manager | Form submission | Approved/denied decision | Email | 1-2 days | No SLA, request sits in inbox |
| 3 | Finance budget check | Finance | Approved request | Budget confirmation | Spreadsheet | 1-3 days | ← RED: No clear budget owner |
| 4 | Legal vendor review | Legal | Vendor MSA | Redlined contract | Email + DocuSign | 2-10 days | ← RED: Legal bottleneck |
| 5 | Contract execution | Ops + Finance | Final contract | Signed agreement | DocuSign | 1 day | Signer unavailable |
| 6 | IT provisioning | IT | Signed contract | Access granted | IT tickets | 1-3 days | Ticket lost or low priority |
| 7 | Confirm with requester | Ops Manager | Access confirmed | Closed loop | Email | 30 min | Often skipped |
## RACI Matrix
| Step | Team Lead | Ops Mgr | Finance | Legal | IT |
|---|---|---|---|---|---|
| 1. Submit request | R/A | I | - | - | - |
| 2. Ops review | C | R/A | I | - | - |
| 3. Budget check | I | C | R/A | - | - |
| 4. Legal review | - | C | C | R/A | - |
| 5. Contract exec | - | A | R | C | - |
| 6. IT provisioning | I | C | - | - | R/A |
| 7. Confirm | C | R/A | - | - | I |
## Bottlenecks
RED — Step 3 (Finance budget check): No named budget owner per department. All requests go to the same Finance inbox.
RED — Step 4 (Legal review): All vendor contracts reviewed sequentially. No template for standard SaaS tools under $10k.
## Recommendations
1. [QUICK] Add department budget owner field to request form — Finance routes immediately
2. [MEDIUM] Create pre-approved vendor template for SaaS tools < $10k/year — bypass full Legal review
3. [STRUCTURAL] Implement procurement tool (Zip, Ramp, Procurify) to track all requests in one place