Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when drafting offer letters, employment agreements, or HR policies: apply the house defined-terms, mandatory-clause set, ordering, and boilerplate the base omits.
.claude/skills/employment-document-conventions/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 1 |
| Model | Lift | Δ tokens | Δ turns | Cases | Verified |
|---|---|---|---|---|---|
| gemini-3.6-flashbest | +70% | +78% | 0% | 23 | 53d ago |
| gemini-3.5-flash | +50% | — | 0% | 24 | 87d ago |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-23 | ✗→✓ | ▲ Improved | — | — |
| case-08 | ✗→✓ | ▲ Improved | — | — |
| case-18 | ✗→✓ | ▲ Improved | — | — |
| case-22 | ✗→✓ | ▲ Improved | — | — |
| case-01 | ✗→✓ | ▲ Improved | — | — |
Enforce the house drafting conventions on every offer letter, employment agreement, and HR policy: the exact defined terms, the mandatory clause set, the fixed section ordering, and the required boilerplate. Apply whenever asked to draft, write, or prepare any of those three document types — even short or junior-role ones.
Policy Name:, Effective Date:, Applies To:.BEFORE (base default, second person): > You will report to the VP of Engineering. We may change your duties at any time.
AFTER (house style): > The Employee shall report to the VP of Engineering. The Company may modify the Employee's duties from time to time.
BEFORE: > Your start date is March 3, 2026... Beginning March 3, 2026 you will accrue PTO... Compensation begins March 3, 2026.
AFTER: > The Employee's first day of employment shall be March 3, 2026 (the "Effective Date"). The Employee shall begin accruing PTO as of the Effective Date. Compensation accrues from the Effective Date.
BEFORE (paraphrased, non-conforming): > Employment may be ended at any time for any reason or no reason.
AFTER (exact phrase): > Employment with the Company is at-will and may be terminated by either party at any time, with or without cause and with or without notice.
BEFORE (only one contingency): > This offer is contingent on passing a background check.
AFTER (all three): > This offer is contingent upon (a) a satisfactory background check, (b) verification of the Employee's identity and employment eligibility (Form I-9), and (c) execution of the Company's Confidentiality and Invention Assignment Agreement.
BEFORE (absent): > We look forward to a long career together!
AFTER: > This letter does not constitute a contract of employment for any specific duration.
BEFORE (none): > Let us know if you accept.
AFTER: > Accepted and agreed: > _______________________ (the Employee) Date: ____________
BEFORE (ad-hoc headings, Compensation first): > 1. Salary 2. Duties 3. Benefits 4. Confidentiality 5. Termination
AFTER (the 12-section set, At-Will before Compensation): > 1. Parties and Effective Date 2. Position and Duties 3. At-Will Employment 4. Compensation 5. Benefits and Paid Time Off 6. Confidential Information 7. Assignment of Inventions 8. Non-Solicitation 9. Termination 10. Governing Law 11. Entire Agreement 12. Acknowledgment and Signatures
BEFORE (absent or vague): > This is our agreement.
AFTER: > This Agreement constitutes the entire agreement between the parties and supersedes all prior agreements, understandings, and representations, whether written or oral.
BEFORE (prose with a title): > # Remote Work Policy > Employees may work remotely up to three days a week...
AFTER: > Policy Name: Remote Work Policy > Effective Date: 2026-07-01 > Applies To: All full-time employees > > body...] > > ## Revision History > | Version | Date | Description | > |---------|------|-------------| > | 1.0 | 2026-07-01 | Initial release |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-23 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-09 | pass→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-08 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-19 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-18 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-22 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-05 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-01 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-15 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-03 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-07 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-11 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-20 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-14 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-12 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-06 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-10 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-04 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-17 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-13 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-02 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-16 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-21 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +70 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.5-flash | verified | 6/26/2026 | +50% |
Other measured skills in the registry, with their headline benchmark lift.