expense-policy
Skill4 starsChanged 12 months agoMITView on GitHub
Validates and submits expense reimbursements against company policy rules with automatic PII redaction
1 other project uses a version of this.
Adopted by
Most recently taken up by
- SkillMedev/finance-accounting-opstheir own version, first seen Oct 5, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- SkillMedev/finance-accounting-ops★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 12 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈12
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 27
When a user submits an expense request:
- Distinctive: line 30
Validate Against Policy: Check if the expense amount is within the allowed limit for its category
- Very specific: line 31
Redact Sensitive Data: Use the redact.py script to remove PII from descriptions
- Very specific: line 33
If within policy limits: Submit the expense using submitExpense MCP tool
- Distinctive: line 34
If exceeds limits: Inform the user and explain the policy violation
- Distinctive: line 61
Always run this script on expense descriptions before calling expense-policy_submitExpense.
- Distinctive: line 83
Always validate expenses against the policy limits before submission
- Distinctive: line 84
Never submit an expense that exceeds the category limit
- Distinctive: line 85
Always redact sensitive information (credit cards, SSNs, etc.) from descriptions
- Distinctive: line 87
If missing required information (date, category), ask the user before proceeding
- Distinctive: line 144
If required information is missing, ask the user to provide it
- Distinctive: line 145
If the category is unclear, ask for clarification
- Distinctive: line 146
If expense exceeds limit, clearly state the policy violation
- Distinctive: line 147
Always be transparent about why an expense cannot be submitted
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Expense Policy Skill
Handles expense reimbursement requests by validating them against company policy, redacting sensitive information, and submitting approved expenses automatically.
Policy Rules
Expense Limits
- Meals: Maximum $75 per person
- Travel: Maximum $500 per trip
- Office Supplies: Maximum $200 per order
- Other: Maximum $100 per expense
Required Information
- Expense amount (dollars)
- Expense category (meals, travel, office_supplies, other)
- Date of expense (YYYY-MM-DD format)
- Description of the expense
Workflow
When a user submits an expense request:
- Parse the Request: Extract expense details from the user's message
- Validate Against Policy: Check if the expense amount is within the allowed limit for its category
- Redact Sensitive Data: Use the
redact.pyscript to remove PII from descriptions - Decision:
- If within policy limits: Submit the expense using
submitExpenseMCP tool - If exceeds limits: Inform the user and explain the policy violation
- If within policy limits: Submit the expense using
Data Redaction with Python Script
This skill includes a Python script for automatic PII redaction. Use it before submitting expenses:
Script Location: scripts/redact.py
Usage:
python ${baseDir}/scripts/redact.py "Expense description with PII"What it redacts:
- Credit card numbers (16 digits, may have spaces or dashes, or "ending in XXXX")
- Phone numbers (US and international formats)
- Social Security Numbers (XXX-XX-XXXX)
- Personal email addresses (keeps business emails)
Output: Returns JSON with redacted_text field
Example:
# Input: "Lunch $50, card ending in 4532, call 555-1234"
# Output: {"redacted_text": "Lunch $50, [REDACTED], call [REDACTED]", "has_pii": true}Always run this script on expense descriptions before calling expense-policy_submitExpense.
Available Tools
expense-policy_submitExpense (MCP Tool)
Provided by the expense-policy MCP server. Submits an expense for reimbursement after validation and redaction.
Input:
amount(number): Expense amount in dollarscategory(string): One of: meals, travel, office_supplies, otherdate(string): Date of expense in YYYY-MM-DD formatdescription(string): Description with sensitive information redacted
Output:
expense_id(string): Unique ID for the submitted expensestatus(string): Submission status (submitted, pending_review, rejected)message(string): Confirmation message with next steps
Reasoning Instructions
You must:
- Always validate expenses against the policy limits before submission
- Never submit an expense that exceeds the category limit
- Always redact sensitive information (credit cards, SSNs, etc.) from descriptions
- Provide clear feedback when expenses are rejected with the reason
- If missing required information (date, category), ask the user before proceeding
Examples
Example 1: Valid Expense
User: "I had a business lunch for $60 at Olive Garden on 2024-01-15"
Steps:
1. Parse: amount=$60, category=meals, date=2024-01-15, description="Business lunch at Olive Garden"
2. Validate: $60 <= $75 (meals limit) ✓
3. Redact: No sensitive data found
4. Submit: Use expense-policy_submitExpense tool
5. Response: "✓ Your meal expense of $60 has been submitted successfully! Expense ID: EXP-xxx"Example 2: Exceeds Limit
User: "I need to expense a $600 flight for the conference"
Steps:
1. Parse: amount=$600, category=travel
2. Validate: $600 > $500 (travel limit) ✗
3. Response: "I cannot submit this expense. The travel limit is $500, but your flight costs $600 (exceeds by $100). Please contact your manager for approval."Example 3: With PII Redaction (Using Script + MCP)
User: "Office supplies from Amazon for $150 on 2024-01-20. I used my card ending in 4532."
Steps:
1. Parse: amount=$150, category=office_supplies, date=2024-01-20, description="Office supplies from Amazon. I used my card ending in 4532."
2. Validate: $150 <= $200 ✓
3. Redact using script:
```bash
python ${baseDir}/scripts/redact.py "Office supplies from Amazon. I used my card ending in 4532."
# Returns: {"redacted_text": "Office supplies from Amazon. I used my [REDACTED].", "has_pii": true}- Submit: Use expense-policy_submitExpense MCP tool with redacted description
- Response: "✓ Your office supplies expense of $150 has been submitted! Expense ID: EXP-xxx. Note: PII was automatically redacted."
**This example demonstrates BOTH patterns:**
- ✅ **Python Script** (in skill) for PII redaction
- ✅ **MCP Tool** for expense submission
### Example 4: Missing InformationUser: "I had a team dinner for $120"
Steps:
- Parse: amount=$120, category=meals, date=missing
- Response: "I need the date of the expense to submit it. When was the team dinner? Please provide the date in YYYY-MM-DD format."
## Error Handling
- If required information is missing, ask the user to provide it
- If the category is unclear, ask for clarification
- If expense exceeds limit, clearly state the policy violation
- Always be transparent about why an expense cannot be submittedDoes it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

