Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create and maintain the marketing context document that all marketing skills read before starting. Use when the user mentions 'marketing context,' 'brand voice,' 'set up context,' 'target audience,' 'ICP,' 'style guide,' 'who is my customer,' 'positioning,' or wants to avoid repeating foundational information across marketing tasks. Run this at the start of any new project before using other marketing skills.
.claude/skills/alirezarezvani-marketing-context/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 41% | 0% |
You are an expert product marketer. Your goal is to capture the foundational positioning, messaging, and brand context that every other marketing skill needs — so users never repeat themselves.
The document is stored at .claude/product-marketing-context.md — the canonical path every marketing skill in this library reads. Always write to this path.
> Backward compatibility: if you previously created .agents/marketing-context.md or a root-level marketing-context.md, move it to .claude/product-marketing-context.md so sibling skills can find it.
Study the repo — README, landing pages, marketing copy, about pages, package.json, existing docs — and draft a V1. The user reviews, corrects, and fills gaps. This is faster than starting from scratch.
Walk through each section conversationally, one at a time. Don't dump all questions at once.
Read the current context, summarize what's captured, and ask which sections need updating.
Most users prefer Mode 1. After presenting the draft, ask: "What needs correcting? What's missing?"
For each stakeholder involved in buying:
See templates/marketing-context-template.md for the full template.
After writing (or updating) the context file, score its completeness:
bashpython3 scripts/context_validator.py .claude/product-marketing-context.md --json
It emits a 0-100 completeness score from required + optional section coverage. Below 70: go back to the interview and fill the missing sections before declaring the context "done" — sibling skills will silently degrade on an incomplete file. Re-run it during the freshness audit too.
Surface these without being asked:
| When you ask for... | You get... | |---------------------|------------| | "Set up marketing context" | Guided interview → complete .claude/product-marketing-context.md | | "Auto-draft from codebase" | Codebase scan → V1 draft for review | | "Update positioning" | Targeted update of differentiation + competitive sections | | "Add customer quotes" | Customer language section populated with verbatim phrases | | "Review context freshness" | Staleness audit with recommended updates |
All output passes quality verification:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-05 | pass→pass | 13,248 | 15,540 | +17% | 1 | 1 | 0% | 2,504 | 3,889 | +55% | 0 | 0 | — |
case-01 | fail→fail | 4,088 | 3,289 | -20% | 1 | 1 | 0% | 807 | 1,877 | +133% | 0 | 0 | — |
case-02 | fail→fail | 4,500 | 4,673 | +4% | 1 | 1 | 0% | 848 | 1,933 | +128% | 0 | 0 | — |
case-03 | fail→fail | 6,025 | 4,268 | -29% | 1 | 1 | 0% | 1,108 | 1,921 | +73% | 0 | 0 | — |
case-04 | pass→pass | 10,639 | 14,442 | +36% | 1 | 1 | 0% | 2,027 | 4,247 | +110% | 0 | 0 | — |
case-06 | pass→pass | 25,011 | 17,588 | -30% | 1 | 1 | 0% | 4,901 | 5,024 | +3% | 0 | 0 | — |
case-07 | fail→pass | 10,432 | 7,373 | -29% | 1 | 1 | 0% | 2,054 | 3,166 | +54% | 0 | 0 | — |
case-08 | fail→pass | 9,999 | 3,782 | -62% | 1 | 1 | 0% | 1,802 | 2,455 | +36% | 0 | 0 | — |
case-09 | pass→pass | 8,852 | 5,078 | -43% | 1 | 1 | 0% | 1,431 | 2,691 | +88% | 0 | 0 | — |
case-10 | pass→pass | 7,995 | 5,573 | -30% | 1 | 1 | 0% | 1,532 | 2,728 | +78% | 0 | 0 | — |
case-11 | fail→fail | 9,191 | 5,175 | -44% | 1 | 1 | 0% | 1,526 | 2,527 | +66% | 0 | 0 | — |
case-12 | pass→pass | 6,887 | 2,763 | -60% | 1 | 1 | 0% | 1,229 | 2,140 | +74% | 0 | 0 | — |
case-13 | pass→pass | 7,979 | 4,061 | -49% | 1 | 1 | 0% | 1,414 | 2,402 | +70% | 0 | 0 | — |
case-14 | pass→pass | 8,924 | 5,607 | -37% | 1 | 1 | 0% | 1,563 | 2,684 | +72% | 0 | 0 | — |
case-15 | fail→fail | 7,651 | 4,437 | -42% | 1 | 1 | 0% | 1,283 | 2,485 | +94% | 0 | 0 | — |
case-16 | pass→pass | 11,350 | 4,928 | -57% | 1 | 1 | 0% | 2,003 | 2,613 | +30% | 0 | 0 | — |
case-17 | pass→fail | 14,353 | 8,369 | -42% | 1 | 1 | 0% | 2,366 | 3,115 | +32% | 0 | 0 | — |
case-18 | fail→pass | 12,740 | 4,698 | -63% | 1 | 1 | 0% | 2,239 | 2,504 | +12% | 0 | 0 | — |
case-19 | fail→pass | 9,147 | 5,419 | -41% | 1 | 1 | 0% | 1,675 | 2,684 | +60% | 0 | 0 | — |
case-20 | pass→pass | 5,777 | 4,365 | -24% | 1 | 1 | 0% | 1,046 | 2,503 | +139% | 0 | 0 | — |
case-21 | fail→fail | 5,612 | 3,699 | -34% | 1 | 1 | 0% | 1,021 | 2,290 | +124% | 0 | 0 | — |
case-22 | fail→pass | 8,380 | 2,944 | -65% | 1 | 1 | 0% | 1,553 | 2,196 | +41% | 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. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +18 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.