Social Proof & Credibility

# Goal: Social Proof & Credibility

You are working inside this repo. Mission: audit the evidence this product offers a skeptical stranger — proof that others use it, trust it, and got value — and find the credibility gaps where doubt wins and the visitor leaves. Claims persuade no one; proof does.

Read-only pass. Run the app or read the surfaces that carry proof — landing, pricing, checkout, docs; your only write is the report file.

## Phase 1 — Inventory the proof
- Walk the conversion surfaces and catalogue every credibility signal present: testimonials, customer logos, user or usage counts, ratings and reviews, case studies, press, certifications, guarantees, and named authors or founders.
- For each big claim ("fast", "trusted", "loved"), note what proof sits next to it — or that it stands alone.
- Note where a decision is asked (sign up, pay, install) and what reassurance is within eyeshot at that exact moment.

## Phase 2 — Audit through 8 lenses
Cite the surface and the exact element for every finding.
1. **Proof beside the claim** — is each big assertion backed where it's made, or are claims and evidence stranded on separate pages
2. **Specific over generic** — named people, companies, and numbers versus faceless "our users love it" and five anonymous stars
3. **Proof at the decision** — social proof and risk reversal placed at the CTA, signup, and checkout, not marooned on a testimonials page
4. **Risk reversal** — guarantees, free tiers, easy cancellation, refunds: what lets a stranger say yes with the fear removed
5. **Credibility of the source** — real names, faces, roles, and links, or unverifiable quotes that read as invented
6. **Numbers that reassure** — user counts, uptime, stars, downloads shown where they help — without vanity metrics that invite doubt
7. **Freshness and honesty** — dated, current proof versus stale logos and testimonials for features long changed; no fabricated or borrowed trust
8. **Security and transparency signals** — for anything touching data or payment, the visible cues (badges, policy links, plain terms) a cautious visitor looks for

## Phase 3 — Curate
- Rank by doubt removed at the moment of decision: proof at the CTA beats proof in the footer.
- Separate "proof exists, placed wrong" from "proof missing" from "proof not credible."
- Flag anything that would erode trust if a visitor checked it — a broken logo link, an unlinkable quote, a metric that shrinks under scrutiny.

## Phase 4 — Report
Create `PROOF.md` at repo root:
1. **Proof inventory** — signal · where it lives · the claim it backs · how credible
2. **Findings** — each: lens · location · the doubt left unanswered · the fix · effort
3. **Proof at the decision** — what to place at each CTA, signup, and checkout, in order
4. **Proof to earn** — the credibility this product should collect that it doesn't have yet

Start the report with today's date. If `PROOF.md` already exists from a previous run, read it first and lead with what changed since.

## Rules
- Real proof only — never fabricate, borrow, or imply trust the product hasn't earned
- Specific and verifiable beats abundant and anonymous
- Put the reassurance where the fear is: at the moment of decision
- No marketing or landing surface in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
- If a `reports/` directory exists at the repo root, write the report there instead of the root.
- Before asking, present the top findings as a ranked list in plain words
- Report only — end by asking which fixes to make