---
card_id: <NN>_sim_<short_slug>
title: "<Card title in plain prose, no Sim-prefix>"
schema_version: 3
estimated_minutes: <int, typically 3-30>
audience_cluster: <clinical | legal | gov | it | research | civic | recovery | builder | new-cluster-with-argument>
mock_ui_theme: <generic_modal | lab_console | voting_portal | claim_console | records_filer | home_hub | hearing_room | new-with-argument>
sources:
  - sources/<NN>_sim_<short_slug>_source.md (if drafting from a source extract)
inspiration: "(optional) Inspired by [Name's] [public work] on [topic]. The framing is theirs; the substrate is ours; the [thing] is yours."
supersedes:
  - (the card this replaces, if any; "remains live until cards_v2 promotes")
---

# <Card title>

> EDIT GUIDE: each `## H2` is one step. The HTML comment immediately
> above each H2 binds the prose to the source step ID. Do not edit
> the comment. Edit the prose freely; the build script round-trips
> it back to the JSON sidecar.

<!-- step.id = landing -->
## <Step 1 title — usually "Landing" or a time-stamped equivalent>

You're <persona> who <inciting context>. <One sentence about the
weight or stakes that brought them to the page.>

The hook on Merkle Trust's landing names the feeling: <hook line in
the persona's language, the question their work was supposed to
have already answered>.

Nothing to type, nothing to commit to. The walk plays itself from
here.

<!-- step.id = install -->
## <Step 2 title — usually "Picking how you'd evaluate">

The walk is the test drive — no install required.

Four real install paths exist when you decide to act. For <persona's
audience cluster>, the order is <cluster-appropriate order>:

**<Path 1 title>.** <2-3 sentences explaining the path's relevance
to this audience.>

**<Path 2 title>.** <2-3 sentences.>

**<Path 3 title>.** <2-3 sentences.>

**<Path 4 title>.** <2-3 sentences.>

A fifth option holds the cursor for now: keep evaluating. The walk
is the test drive.

<!-- step.id = first_concrete -->
## <Step 3 title — the dramatic moment>

A sandboxed Merkle Trust loads with a synthetic case: <named
scenario>. The <robot persona> on the desk animation places the
material under the seal one piece at a time.

<List the synthetic data items the sim loads, with brief context
for each.>

The walk takes the <picked sidebar scenario> — <one sentence
explaining why this scenario is the moment that matters most>.

<2-4 paragraphs of the dramatic moment playing out. The robot
fetches relevant items. The ticker streams the SHAs. The substrate
demonstrates the capability that makes this card defensible.>

<If the moment earns an ASCII slab, place it here as a fenced code
block. Slabs are appropriate for: court records, deposition
transcripts, technical diagrams, recall letters, contribution-flow
diagrams, etc. Skip the slab if it would feel forced.>

<!-- step.id = is_this_real -->
## <Step 4 title — usually "\"Is this real?\"">

<Three plain-English bullets, each bolded and explained in 2-3
sentences. The bullets address the visitor's most likely
skepticism.>

<Optional paragraph: cluster-specific posture (privilege for legal,
sunshine-law for gov, FERPA for school, GDPR for consumer, etc.).>

The "Anchored here" link resolves to a real receipt: a real merkle
root, a real timestamp, a one-line explanation.

The **.md** button at the bottom puts a structured summary of <this
card's core pattern> into your tag-along bundle. If you would like
a critical reviewer to read it and ask questions <a relevant
counterparty> might ask, the comment field carries one to your own
claude.ai session — with the full prompt shown to you before it
sends.

<!-- step.id = ceremony -->
## <Step 5 title — usually "The ceremony moment">

Run a ceremony. Fifteen seconds. It's the thing that makes
everything you just saw defensible.

The <robot persona> presses the <appropriate stamp/seal>. The
ticker streams the hex of each leaf as it joins today's tree. Real
SHA-256 fires in your browser, real merkle leaves combine into a
real root, the new anchor lands at the configured chain. Progress
bar reads "done — <count> files attested, <count> critical files
verified, new anchor at " followed by the first eight hex
characters of the root.

The output: <one sentence in the persona's register about what just
ran>. <Optional second sentence about the cluster-specific
implication.>

<!-- step.id = close -->
## <Step 6 title — usually "The close">

Three ways to leave, no pressure either way.

Whichever direction the walk takes from here, you take a signed
package with you — yours to keep, downloaded directly when you
finish. The package contains the manifest with the merkle root of
every file, a certificate of registration carrying your earliest
attestation timestamp, the architecture docs, the verification
scripts (`verify.sh` and `verify.py`, stdlib only), a
`claude_prompt.md` you can paste into any LLM for an independent
audit, the JSON schemas, and a `DISCLAIMERS.md`.

A second file rides along with the cert: a recovery seed. Twelve
words written into a file with a self-protecting preamble — if it
ever ends up pasted into an LLM by accident, the preamble tells the
LLM to STOP and warn the user. Daily access uses the cert; the seed
is the fire exit.

The most useful close for a <persona> is <cluster-specific kit>.
<One or two sentences naming the partner kit and the next concrete
opportunity to use it.>

<!-- finish_text -->
# Finish text

That was the simulated path through <named scenario>. The full card
breaks out the savings, <card-specific topic>, and a <prediction
shape> that's yours to test.

<!-- voice_check -->
> Voice check before submission:
> - TO the visitor / ABOUT named persona in third person
> - No "the visitor sees" / "this card surfaces"
> - No "Picture this:" / catastrophe-framing
> - No italic markdown
> - Bold OK on bullet labels and button names
> - Inline quotes use double quotes
> - No "patent pending" / "revolutionary blockchain"
> - Truth claims map to real components or are framed as build-on
> - ASCII slab earns its place or is omitted