--- AUDIT RESULTS 📊 Quality Score: 30% (Under 50%: Needs significant work | 50-74%: Basic level, gaps remain | 75-89%: Good, minor gaps | 90-100%: Production-ready) Reason: The prompt states a platform and industry but omits audience, offer, constraints, format, and claims handling, leaving most execution decisions to guesswork. 1. Components: - Goal: Partial — "write captions" is the action, but the underlying objective (drive engagement, sales, follows, sign-ups) isn't specified. - Context: Partial — platform (Instagram) and industry (skincare) are given; brand voice is only loosely implied by "fun and engaging." - Core Message or Offer: Missing — no specific product, benefit, or promotion to feature. - Constraints: Missing — no length limits, required disclaimers, or words/claims to avoid. - Role/Audience: Missing — no target demographic or segment named. - Output Format: Missing — number of captions, caption+hashtag structure, or variant count not specified. - Technical Specs: Missing — no character limits or Instagram-specific formatting notes (line breaks, emoji use, hashtag count). - Claims & Substantiation Handling: Missing — no instruction on how to handle skincare efficacy claims or testimonials. 2. 🔒 Safety & Structure: - Protection against prompt-behavior alteration: Not Applicable — the submitted text contains no instructions attempting to redirect model behavior. - Personal/confidential data exposure: Not Applicable — no PII appears in the prompt. - Unsubstantiated claims risk: Present — skincare copy generated without claims guidance risks the model inventing specific efficacy statistics or testimonials. - Regulated-claims risk left unflagged: Present — skincare is a regulated category (cosmetic/health claims), and the prompt gives no instruction to avoid or flag medical/comparative claims. - Third-party brand/trademark reproduction risk: Not Applicable — no competitor or existing campaign is referenced. 3. ⚠️ Risks Found: - Risk: Fabricated product claims. Cause: no claims-handling instruction. Impact: captions could include invented statistics ("clinically proven to reduce wrinkles by 40%") that create regulatory/legal exposure. - Risk: Unusable output format. Cause: no character limit, hashtag count, or variant number specified. Impact: user may receive captions too long for Instagram or the wrong quantity for their needs. - Risk: Generic, unfocused messaging. Cause: no target audience defined. Impact: captions may not resonate with the brand's actual customer segment. - Risk: Missed regulatory review trigger. Cause: skincare/health category claims left unflagged. Impact: copy could inadvertently include claims requiring substantiation or legal review before publishing. 4. 💡 Recommendations: 🚨 Critical: Add explicit instruction to avoid fabricating specific claims, statistics, or testimonials; flag any claim needing user-supplied substantiation. 🚨 Critical: Specify target audience/segment to focus tone and messaging. 🔴 High: Specify output format — number of caption variants, caption+hashtag structure. 🔴 High: State the core message or offer the captions should center on (new product, general brand awareness, seasonal promo, etc.). 🟠 Medium: Add Instagram-specific technical specs (character guidance, hashtag count, emoji preference). 🟡 Low: Clarify brand voice beyond "fun and engaging" (e.g., playful, minimalist, science-forward). 5. ✅ Improved Prompt: Write 5 Instagram captions for (assumption: a skincare brand — name and specific product not provided by user) targeting (assumption: general skincare-interested Instagram audience, since no specific demographic was given). Each caption should be fun and engaging in tone, playful but not silly, and centered on (assumption: general brand awareness and product benefits, since no specific offer or product was named). Keep each caption under 150 characters for the main text, followed by 3-5 relevant hashtags. Do not invent specific statistics, clinical claims, percentages, or testimonials about the product's effectiveness — if a claim would require substantiation (e.g., "reduces wrinkles," "clinically proven," "dermatologist recommended"), flag it explicitly as an assumption needing user-supplied evidence rather than stating it as fact. Avoid comparing the brand to any named competitor. Provide the 5 captions as a numbered list, each with its own hashtag set. 6. Score Comparison: Original Score: 30% — missing audience, offer, format, constraints, and claims handling leaves most execution choices unspecified. Improved Score: 82% — adds audience assumption, output format, technical specs, and claims-handling guardrails; still capped below 90% because the actual product name, specific offer, and true target demographic remain assumptions rather than user-confirmed facts. Improvement: +52 points --- ℹ️ Note: This tool optimizes the initial prompt — it doesn't verify factual, legal, regulatory, or financial accuracy of any claim referenced; final decisions remain the user's responsibility. AI is a probabilistic system, so wording and scores may vary slightly between runs. We recommend a final legal/compliance check before professional or commercial use. Built for Claude Sonnet 5; other AI models, providers, and versions are untested. Use at your own discretion. This tool is not intended as a sole safeguard against attempts to alter its behavior or against advertising/regulatory compliance review. 💡 Don't have a prompt yet? Search "Marketing Request Auditor" on PromptBase to turn a rough client brief into a clear one, or "Marketing Prompt Generator" to build a ready-to-use prompt from scratch — then bring it back here to audit and refine. ❓ A few quick questions that could improve the next pass: 1. What's the actual brand name and specific product/offer you want featured in these captions? 2. Who is your target audience (age range, skin concerns, lifestyle)? 3. Do you have any real product claims, ingredients, or testimonials you'd like included, and are there any words or claims you need to avoid?