Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write a professional press release for any announcement. Use when asked to write a press release, media announcement, news release, or press statement. Produces a structured press release with headline, dateline, body, boilerplate, and media contact — ready to send to journalists.
.claude/skills/itamarzand88-press-release/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 73% | 0% |
<!-- source: press-release-writer — https://raw.githubusercontent.com/mohitagw15856/pm-claude-skills/main/skills/press-release/SKILL.md -->
Writes press releases that journalists actually read — structured around the news angle, not the desire to promote.
FOR IMMEDIATE RELEASE / EMBARGOED UNTIL: Date and time]
City, Date] — Opening paragraph: Who, What, When, Where, Why in 2-3 sentences. A journalist should be able to run this paragraph alone. No background, no context, no company history.]
Second paragraph: the significance. Why does this matter? What does it mean for customers or the industry?]
Third paragraph: quote from executive. Human and specific. Not a restatement of the headline.]
"Quote text — specific, adds something the facts do not say]," said Name], Title] at Company]. "Second sentence extending the thought]."
Fourth paragraph: supporting detail — data, customer names with permission, additional context]
Fifth paragraph optional: what happens next, when it goes live, what people can do]
ENDS
Notes to editors:
About Company] Boilerplate: 3-4 sentences. What the company does, when founded, where based, key facts. Factual not promotional.]
Media contact: Name] | Title] | Email] | Phone] | Hours/timezone]
Would a journalist care? Is the headline the full story? Is there a human angle? Is the quote something a human would say? Can the first paragraph stand alone?
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,536 | 9,908 | +4% | 1 | 1 | 0% | 1,745 | 2,670 | +53% | 0 | 0 | — |
case-02 | fail→pass | 11,072 | 8,112 | -27% | 1 | 1 | 0% | 2,065 | 2,536 | +23% | 0 | 0 | — |
case-03 | fail→pass | 10,652 | 9,750 | -8% | 1 | 1 | 0% | 1,962 | 2,766 | +41% | 0 | 0 | — |
case-04 | fail→pass | 8,815 | 8,976 | +2% | 1 | 1 | 0% | 1,714 | 2,501 | +46% | 0 | 0 | — |
case-05 | fail→pass | 7,228 | 8,211 | +14% | 1 | 1 | 0% | 1,411 | 2,448 | +73% | 0 | 0 | — |
case-06 | pass→pass | 7,068 | 6,731 | -5% | 1 | 1 | 0% | 1,531 | 2,213 | +45% | 0 | 0 | — |
case-07 | fail→pass | 8,634 | 7,761 | -10% | 1 | 1 | 0% | 1,707 | 2,426 | +42% | 0 | 0 | — |
case-08 | pass→pass | 8,852 | 9,386 | +6% | 1 | 1 | 0% | 1,811 | 2,747 | +52% | 0 | 0 | — |
case-09 | fail→pass | 9,386 | 7,039 | -25% | 1 | 1 | 0% | 1,996 | 2,226 | +12% | 0 | 0 | — |
case-10 | pass→pass | 8,263 | 6,926 | -16% | 1 | 1 | 0% | 1,771 | 2,257 | +27% | 0 | 0 | — |
case-11 | fail→pass | 7,477 | 6,818 | -9% | 1 | 1 | 0% | 1,473 | 2,327 | +58% | 0 | 0 | — |
case-12 | pass→pass | 10,202 | 7,777 | -24% | 1 | 1 | 0% | 1,964 | 2,568 | +31% | 0 | 0 | — |
case-18 | pass→pass | 10,154 | 6,582 | -35% | 1 | 1 | 0% | 2,091 | 1,956 | -6% | 0 | 0 | — |
case-13 | pass→pass | 12,233 | 9,696 | -21% | 1 | 1 | 0% | 2,685 | 2,938 | +9% | 0 | 0 | — |
case-14 | pass→pass | 7,722 | 6,326 | -18% | 1 | 1 | 0% | 1,748 | 2,081 | +19% | 0 | 0 | — |
case-15 | fail→pass | 7,194 | 6,683 | -7% | 1 | 1 | 0% | 1,427 | 2,243 | +57% | 0 | 0 | — |
case-16 | fail→pass | 9,364 | 8,105 | -13% | 1 | 1 | 0% | 1,928 | 2,469 | +28% | 0 | 0 | — |
case-17 | pass→pass | 7,310 | 8,921 | +22% | 1 | 1 | 0% | 1,427 | 2,580 | +81% | 0 | 0 | — |
case-19 | pass→pass | 9,577 | 8,202 | -14% | 1 | 1 | 0% | 1,926 | 2,467 | +28% | 0 | 0 | — |
case-20 | pass→pass | 10,500 | 7,974 | -24% | 1 | 1 | 0% | 2,038 | 2,398 | +18% | 0 | 0 | — |
case-21 | fail→pass | 8,789 | 5,837 | -34% | 1 | 1 | 0% | 1,790 | 2,112 | +18% | 0 | 0 | — |
case-22 | pass→pass | 8,398 | 6,401 | -24% | 1 | 1 | 0% | 1,676 | 1,970 | +18% | 0 | 0 | — |
case-23 | pass→pass | 5,133 | 5,705 | +11% | 1 | 1 | 0% | 1,096 | 1,760 | +61% | 0 | 0 | — |
case-24 | pass→pass | 13,600 | 12,891 | -5% | 1 | 1 | 0% | 2,702 | 3,330 | +23% | 0 | 0 | — |
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. 24 cases were attempted. The headline lift of +46 percentage points is the difference between those two pass rates over the 24 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.