borrow-integration-review

Borrow Integration Review (Blue & Midnight)

An orchestrated review of a borrow integration against the Morpho Integrator UX Playbook. You are the orchestrator: the per-foundation compliance agents do the check-level work — each agent carries its own checks and fail signals — and report back; you aggregate their verdicts into the rubric checklist, judge red flags, and deliver one report.

Baseline checklist: docs/rubrics.md — read it first. For borrow, the applicable rows are the Shared table plus the Variable Rate Borrow and/or Fixed Rate Borrow tables, depending on which rate types the integration offers. Every applicable row appears in the final report, cited by its criterion (e.g. "fails Origination fee shown as your fee, separate").

Procedure

1. Scope the review. Identify what's being reviewed: a codebase (find the UI components, copy, quoting and transaction logic), screenshots, or a written flow description. Determine which rate types the integration offers — variable (Blue), fixed (Midnight), or both — and collect the concrete artifact paths; the agents need both.
2. Spawn all seven compliance agents in one message so they run in parallel. Each gets: the artifact paths/context, the rate type(s) — "variable-rate borrow (Blue)", "fixed-rate borrow (Midnight)", or both — and the instruction to run its checks and report verdicts.
   • morpho-integration:vocabulary-compliance — product vocabulary
   • morpho-integration:attribution-compliance — Powered by Morpho badge
   • morpho-integration:disclosure-compliance — disclosure gate
   • morpho-integration:rate-transparency-compliance — rate labelling and fixed/variable distinction, fee separation, orderbook quoting
   • morpho-integration:conversion-compliance — bundled borrow, live preview, input affordances, manage flows
   • morpho-integration:clarity-safety-compliance — recap order, tooltips, warnings, live health/LTV/liquidation, market transparency, safe default LTV, maturity display and obligation, required collateral
   • morpho-integration:discoverability-compliance — value prop, nudges, integrated position
3. While agents run, do the red-flag pass yourself — holistic judgments that need the whole flow in view, not a single row. Flag the flow as weak if any of these are true:
   • More than one signature for a single borrow.
   • The headline rate is below the fold or smaller than secondary details.
   • No live preview of the outcome or health as the user types.
   • Risk (LTV, liquidation, low liquidity, maturity) surfaced only after confirmation.
   • A rate whose fixed-vs-variable nature is ambiguous on any screen.
   • A fixed-rate quote that doesn't move when the borrow amount changes, or isn't refreshed before signing.
   • Protocol jargon (LLTV, shares, IRM, orderbook mechanics, oracle scaling) in the primary flow instead of behind a tooltip or advanced reveal.
   • An eligible-asset prompt that doesn't lead to a CTA.
   • More than a tap or two to learn what market the user is borrowing from.
4. Aggregate. Match each agent's named checks to the rubric rows in docs/rubrics.md — the names correspond one-to-one. Every applicable row must land a verdict; a row no agent could assess is UNVERIFIED, never guessed. If an agent contradicts your own observation, re-check the evidence — every verdict must be traceable to a file, screenshot, or quoted flow step.
5. Report in this exact structure — every Shared row plus the Variable and/or Fixed rows for the rate types offered appears once, split into the playbook's three goals using the rubric's Goal column (Compliance → Compliant; Conversion → Converts, grows TVL; Clarity and Discoverability → Smooth & discoverable), keeping rubric order within each group:

# Borrow Integration Review

## Verdict
<one paragraph: overall state; rate types covered; count of Critical / Recommended / Nice-to-have failures>

## Compliant
| Criterion | Priority | Verdict | Evidence / Fix |
| --- | --- | --- | --- |
| Approved vocabulary | Critical | PASS / FAIL / UNVERIFIED / N-A | <evidence for the verdict; for FAIL, the shortest fix> |
| ... every Compliance-goal row ... | | | |

## Converts, grows TVL
| Criterion | Priority | Verdict | Evidence / Fix |
| --- | --- | --- | --- |
| ... every Conversion-goal row ... | | | |

## Smooth & discoverable
| Criterion | Priority | Verdict | Evidence / Fix |
| --- | --- | --- | --- |
| ... every Clarity- and Discoverability-goal row ... | | | |

## Red flags
<any triggered flow red flags, with evidence>

## Launch self-review
<the checklist below, checked off from the findings>

Rules: every verdict needs evidence (full file path, screen, or quoted description); a single Critical failure means the integration is not launch-ready — say so plainly in the Verdict paragraph.

1. Launch self-review, checked off from the findings (skip lines for a rate type the integration doesn't offer):
   • Is the language aligned (no staking / investment / guaranteed / risk-free)?
   • Is "Powered by Morpho" visible wherever users interact with Morpho functionality?
   • Is there a distinct moment where users acknowledge Morpho's disclaimer?
   • Is every rate clearly labelled variable or fixed, and are the two products distinguishable?
   • Is the rate transparent about its components, with the integrator's fee separated?
   • Can a user name the market they're using without leaving the flow?
   • Are health, LTV, and liquidation shown live and before confirmation?
   • Fixed: is the maturity and its consequence unmistakable?
   • Fixed: is the quote requoted on input change and re-simulated before signature?
   • Fixed: is the required collateral computed and shown from the borrow amount?
   • Does borrow complete in a single signature where the protocol allows?
   • Do eligible-asset prompts actually lead somewhere?