Back to search
service-office-skills
SkillWorks with Claude CodeGitHub
Four Claude skills for a service company's office: deposit matching, daily jobs review, financing fee check and a routine watchdog. Read-only, exact to the cent, tested.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/chrstphrpond--service-office-skills)HTML
<a href="/listings/chrstphrpond--service-office-skills"><img src="/api/listings/chrstphrpond--service-office-skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
i
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Copy a skill's folder into `~/.claude/skills/`.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 49 lines
- # Service office skills
- Four Claude skills for the office of an HVAC and plumbing company. Each one takes a daily or monthly check that someone does by hand, runs it the same way every time, and lists only what a person has to look at.
- Written by Christopher Pond Maquidato in November 2026. They run on made-up sample data. The column names, bank wording and fee rates are placeholders until they are set from real exports.
- | Skill | The question it answers | Reads |
- |---|---|---|
- | [`deposit-match`](deposit-match/) | Did every deposit reach the bank, to the cent? | Payments export, bank export |
- | [`jobs-review`](jobs-review/) | What was wrong with yesterday's jobs? | Completed jobs export |
- | [`financing-fees`](financing-fees/) | Did the financing company deduct the right fee on each loan? | Funding statement, payments export |
- | [`routine-watchdog`](routine-watchdog/) | Did every routine run today, and if not, why? | Run log, schedule |
- 
- One of the four reports. Each skill's folder shows its own.
- ## How they are built
- The four share the same rules.
- - **A script does the work, not the model.** Matching, adding up and deciding a status are done by plain Python in whole cents. Claude runs the script and reports what it found.
- - **Read-only.** No skill posts, edits, marks or restarts anything. Each one says what a person should do and stops.
- - **Exact or nothing.** There is no tolerance and no best guess. One cent off, two possible answers or a missing rate goes to a person with the numbers shown.
- - **Bad input stops the run.** A missing column, a bad amount or a duplicate ID writes no report at all.
- - **The totals are proven.** Each script checks that its numbers add up before it writes a report, and refuses to write one if they don't.
- - **Three exit codes.** 0 is clean, 1 means something needs a person, 2 means the input couldn't be read. A scheduler can alert on 1 and 2.
- Each folder has a `SKILL.md` that Claude reads, the script, a config file, a sample week with the report it produces, and tests.
- ## Try one
- ```
- cd deposit-match
- python3 scripts/match_deposits.py --payments sample-data/payments.csv --bank sample-data/bank.csv --config config.example.json --out /tmp/deposit-match-try
- python3 -m unittest discover -s tests
- ```
- Python 3.9 or newer. Nothing to install.
- ## Install as skills
- Copy a skill's folder into `~/.claude/skills/`.
- ## Not done yet
- - Reading ServiceTitan, QuickBooks Online or the bank directly. Every skill works from an export or a log file.
- - Sending the alert. The exit codes are there for it; email or chat delivery is not built.
- - The web dashboard that would show each person their own list. Every skill writes a `report.json` for it.
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.