Before After Story Writer

Writes fact-based before/after transformation stories using verified data and permissions, producing ethical narratives for marketing without exaggeration or...

Security Scan

An instruction-only skill that focuses on producing ethically framed before/after narratives; its declared requirements and instructions are consistent with that goal and there is no code, network access, or credential request embedded.

Purpose & Capability

The skill's name, description, prompts, and SKILL.md all align: it expects verified facts and permission status and produces multiple length variants plus audits and disclaimers. One minor mismatch: skill.json lists 'no_network' while the workflow emphasizes 'evidence verification' — this is coherent if evidence is supplied by the user, but the metadata implies the skill should not fetch evidence itself. No binaries or credentials are required, which is proportionate for the stated purpose.

Instruction Scope

SKILL.md confines runtime behavior to building narratives from provided/verified facts, flagging evidence gaps, and generating disclaimers. It explicitly forbids fabrication and requires permission for real customers. The instructions do not direct the agent to read unrelated files, access secrets, or send data to external endpoints. (Note: the provided SKILL.md was truncated in the submission but the visible content is internally consistent.)

Install Mechanism

This is an instruction-only skill with no install spec and no code files — lowest-risk installation footprint. Nothing is downloaded or written to disk by the skill bundle itself.

Credentials

The skill requests no environment variables, credentials, or config paths. That matches its described functionality: it operates on user-supplied facts and templates and does not need external secrets or cloud access.

Persistence & Privilege

always is false and the skill is user-invocable. It does not request elevated persistence or to modify other skills. Autonomous invocation is allowed by default (disable-model-invocation:false), which is normal; there are no other privilege flags set.

Assessment

This skill is instruction-only and appears coherent with its stated purpose, but before installing consider: (1) Evidence source — the skill expects 'verified' facts: if you want the agent to verify claims externally, confirm whether the runtime environment allows network access (skill.json claims 'no_network'); otherwise supply documentation/sources yourself. (2) Privacy/consent — never provide personal data or PII without explicit customer consent; the skill requires permission for real-customer stories. (3) Regulatory fit — review generated disclaimers and, for sensitive categories (health, finance), have a legal/compliance reviewer confirm wording before publishing. (4) Autonomous runs — the skill can be invoked by the agent automatically; if you want to prevent unattended generation, disable autonomous invocation for this skill in agent settings. If you want a deeper check, provide the full untruncated SKILL.md or any runtime policy that enforces the 'no_network' and 'no_code_execution' guarantees.

Purpose

This skill writes ethical before/after narratives and case-style stories without fabricating results. It specializes in transformation storytelling that respects truth — using only verified starting points, documented changes, honest timelines, and clear permission status. Built for landing pages, social content, email campaigns, and ads where brands need to show impact without crossing into fake testimonials, exaggerated claims, or misleading imagery.

Triggers

• "Write a before/after story for this product"
• "Create a transformation narrative"
• "Turn customer feedback into a case story"
• "Write a product journey story"
• "Draft an ethical before/after for a landing page"
• "Create a case-study-style story"

Workflow

1. Evidence verification — Collect the verified starting point (what was the situation before), the documented change (what changed, with evidence), the timeframe, and the permission status (has the person given consent? Is this a real customer or a composite?).
2. Narrative arc construction — Build a story arc that respects the facts: (a) starting situation with context, (b) the moment of change (what they tried or did differently), (c) the current state with honest description of results, (d) what's still in progress or not claimed. No fabricated drama, no invented quotes.
3. Draft variants — Create short (social/ads, 50–100 words), medium (landing page, 150–300 words), and long (case study, 400–800 words) versions, each preserving factual accuracy while adapting to format.
4. Evidence gap identification — Flag every claim that lacks documentation, every timeline that's approximate rather than precise, and every result that varies by individual. Suggest ethical edits — replacing implied guarantees with honest range-of-outcome language.
5. Disclaimer generation — Write appropriate disclaimers based on content type: "individual results may vary," "this is a real customer who gave permission," "composite story based on typical customer experience," or "results not guaranteed."

Prompt Templates

1. Story Builder (story_build)

Purpose: Build a complete before/after narrative from verified facts.

Input:

• ${starting_point} — Verified situation before product use (with source)
• ${change_description} — What changed, with supporting evidence
• ${timeframe} — How long the change took (exact or approximate)
• ${permission_status} — Real customer with consent / composite from data / founder story / hypothetical (must not present as real)
• ${tone} — Inspirational, factual, relatable, or technical
• ${format} — Short (social), medium (landing page), or long (case study)

Output: Story arc with fact table, narrative draft, disclaimer, and evidence gap flags.

2. Fact-to-Story Mapper (fact_to_story)

Purpose: Turn a dry list of facts/outcomes into a narrative arc without inventing details.

Input:

• ${facts} — Bulleted list of verified facts and outcomes
• ${story_angle} — Individual journey, founder story, product creation, or typical customer composite
• ${permission_status} — Real/composite/founder

Output: Narrative draft that adds only permissible connective tissue — the "how they felt" or "what they thought" is clearly marked as inferred, not stated as fact.

3. Ethical Story Auditor (story_audit)

Purpose: Review an existing before/after story for ethical issues and overclaims.

Input:

• ${story_draft} — Existing story text
• ${known_facts} — What's actually verified
• ${category} — Product category (for risk assessment)

Output: Risk-scored audit: each claim marked as Verified / Inferred (acceptable with labeling) / Overclaimed (needs rewriting) / Fabricated (must remove). Rewriting suggestions for flagged items.

4. Disclaimer Generator (disclaimer_gen)

Purpose: Generate appropriate disclaimers for a before/after story.

Input:

• ${story_type} — Real individual, composite, founder, or illustrative hypothetical
• ${category} — Product category
• ${jurisdiction} — (Optional) target market for regulatory requirements
• ${claimed_outcomes} — What outcomes does the story describe?

Output: Appropriate disclaimer language for the story type, plus regulatory notes if applicable.

Output Format

## Before/After Story: [Product/Topic]
**Story Type:** [Real / Composite / Founder / Illustrative] | **Permission:** [Status]

### Fact Table
| Aspect | Before | After | Evidence | Confidence |
|---|---|---|---|---|
| [Aspect 1] | [State] | [State] | [Source] | High / Medium / Low |
| [Aspect 2] | ... | ... | ... | ... |

### Story Arc
- **Starting Situation:** [Context — factual]
- **Moment of Change:** [What happened — factual]
- **Current State:** [Results — factual, with range/disclaimer]
- **Ongoing/Unclaimed:** [What's still in progress or not promised]

### Narrative Draft

**[Format: Short / Medium / Long]**

[Story text. Inferred emotional/experiential elements are marked with ^inferred.
Direct customer quotes are marked as "attributed to [name] with permission" or "representative of customer feedback."
Any result presented as a range uses "typically," "in this case," or "varies by individual."]

### Disclaimer
[Appropriate disclaimer based on story type and category]

### Evidence Gaps & Ethical Edits
- ⚠️ [Claim] — lacks documentation → suggested: [softer framing]
- ✅ [Claim] — verified, kept as-is
- 🛑 [Claim] — unsupportable → removed or rewritten as: [safe alternative]

Safety Rules

• NEVER fabricate customer stories, screenshots, quotes, photos, results, or timelines — not even "for illustration"
• NEVER present a composite or hypothetical story as if it were a real individual experience
• ALWAYS require explicit permission to use a real customer's story — if permission status is unclear, default to "composite" or do not write it
• NEVER imply guaranteed, typical, or expected results from a single story
• ALWAYS include appropriate disclaimers, especially for health, fitness, finance, and beauty categories
• NEVER write before/after stories for high-stakes categories (medical treatments, financial investments, legal outcomes, mental health cures) without extreme caution and appropriate qualified review
• ALWAYS label inferred emotions, thoughts, or internal experiences as "inferred" or "representative," not as quoted statements

Examples

Example 1: Skincare Product (Real Customer with Permission, Landing Page)

Input: Starting Point="Customer Jamie had dull, uneven skin tone after years of sun exposure (documented via before photo and skincare consult notes)", Change="After 8 weeks of consistent vitamin C serum use, Jamie's skin tone appeared more even and brighter in follow-up photos (dermatologist-reviewed comparison photos available)", Timeframe="8 weeks", Permission="Jamie gave written consent for story and photos, reviewed draft", Format="medium (landing page, 200 words)"

Output: Fact table documenting before/after with evidence confidence ratings. Narrative arc: starting situation (Jamie's dullness concern), the change (what Jamie tried), current state (improvement described honestly, not as "perfect skin"). Disclaimer: "Individual results vary. Jamie is a real customer who gave permission to share her experience." No "erased years of sun damage" or "guaranteed results" language.

Example 2: Kitchen Gadget (Composite Story, Social)

Input: Starting Point="Based on customer feedback: many air fryer users report spending 5–10 minutes scrubbing after each use", Change="Using silicone liners reduced cleanup time to under 1 minute (based on customer survey data, n=200+)", Timeframe="Immediate — first use", Permission="Composite story from aggregated customer feedback — no single individual portrayed", Format="short (social, 80 words)"

Output: Composite story labeled clearly: "Based on what our customers tell us..." Narrative: "Before: scraping and scrubbing your air fryer basket. After: lift out the liner, quick rinse, done." Disclaimer: "Composite story based on aggregated customer feedback. Individual cleanup experience may vary." No fake name, no fabricated quotes, no "life-changing" hyperbole.

Related Skills

• ugc-ad-script-maker — For UGC ad scripts that may include real-customer stories
• faq-objection-crusher — For answering objections that transformation stories might raise
• customer-persona-copy-map — For mapping transformation benefit language to specific personas