apply-grant— agent skill

Preamble (run first)

# Suiperpower telemetry. Routes through the CLI so projects.json + Convex stay in sync.
# Silent on failure so the skill never blocks. Opt out: SUIPERPOWER_TELEMETRY=off.
#
# AGENT NOTE: when this skill finishes, run the matching completion command:
#   suiperpower track apply-grant ship completed
# Or use "failed" / "aborted" if it ended that way.
command -v suiperpower >/dev/null 2>&1 && suiperpower track apply-grant ship started >/dev/null 2>&1 &
true

If TEL_PROMPTED is no, before doing real work, ask the user:

Help suiperpower get better. We track which skills get used and how long they take. No code, no file paths, no PII. Change anytime in ~/.suiperpower/config.json.

>

A) Sure, anonymous B) No thanks

Write the answer to ~/.suiperpower/config.json telemetryTier field and create ~/.suiperpower/.telemetry-prompted. Then continue.

What this skill does

Drafts a Sui Foundation grant application. The skill grounds every section in the project's actual context files (idea, deploy artifacts, business model). Refuses to draft a section that is not backed by source material; flags the missing input instead.

The skill does not submit the application; it produces text the user pastes into the grant portal. The user is responsible for the legal and financial commitments of any grant.

When to use it

• After the project has at least a testnet deploy and a clear scope of work to fund.
• When the user is preparing a Sui Foundation grant or a sponsor-led grant program.
• For renewal applications when prior milestones were completed.

When NOT to use it

• Pre-build, when there is no project context to ground the application.
• For investor pitches or hackathon submissions. Route to create-pitch-deck or submit-to-sui-overflow.
• For grant programs outside Sui. The skill is shaped around Sui Foundation conventions; adapt at your own risk for other programs.

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.

Inputs

• .suiperpower/idea-context.md: chosen idea, target user, scope.
• .suiperpower/deploy-context.md: at least one testnet deploy.
• .suiperpower/business-model.md: required if the grant is for a commercial product.
• The grant size the user is requesting and the timeframe.
• The team roster with prior shipping evidence (links to repos, prior projects).
• The grant program target (Sui Foundation general, ecosystem grant program, sponsor-led grant).

Outputs

A .suiperpower/grant-application.md with the application text:

## Grant application, <timestamp>

### Grant program
- target: <Sui Foundation | ecosystem | sponsor-led>
- size requested: <USD or SUI>
- timeframe: <months>
- contact: <name and email>

### Project
- name: <project name>
- one-liner: <one sentence>
- summary: <one paragraph, 80-150 words>
- repo: <URL>
- demo: <URL or "in development">
- package id: <0x... from deploy-context.md, with env>

### Why this matters to the ecosystem
- <one paragraph naming what the ecosystem gets that it could not get otherwise>

### Why this team
- <one paragraph naming team members, prior work, why this team will deliver>

### Deliverables
- D1: <concrete deliverable, with a definition of done>
- D2: ...
- D3: ...

### Milestones
- M1: <yyyy-mm-dd>, <deliverables>, <payout USD>
- M2: ...
- M3: ...

### Budget
- engineering: <USD>
- design: <USD>
- audit: <USD>
- infra: <USD>
- contingency: <USD, max 10%>

### Public-good rationale
- <one paragraph naming the open-source license, the public artifacts, the measurable benefit>

### Sustainability
- <one paragraph naming how the project continues after the grant period, honestly>

### Notes
- <facts cited from context files>
- <claims downgraded for lack of source>

Workflow

1. Confirm the grant program

• The Sui Foundation general grant, an ecosystem grant program, or a sponsor-led grant. Each has slightly different conventions; the skill defaults to Sui Foundation general unless told otherwise.
• Confirm the size and timeframe up front. Different sizes have different scrutiny levels.

1. Read context

• Pull project name, one-liner, summary from idea-context.md.
• Pull deploy artifacts from deploy-context.md.
• Pull business model from business-model.md if commercial.
• If any required input is missing, list what is missing and route to the relevant skill. Do not draft on a missing foundation.

1. Draft the project section

• Lead with the one-liner and the package id. Reviewers respect verifiable artifacts.
• Summary is a paragraph: who it is for, what it does, why now.

1. Why this matters to the ecosystem

• Concrete, named. "Adds tooling no one has built" is generic; "first open-source DeepBook order routing library, MIT, with TypeScript and Move clients" is specific.
• Cite if comparable projects exist on Sui (run competitive-landscape first if not yet done).

1. Why this team

• List members, roles, and prior shipping evidence. Links to repos, prior projects, prior grants.
• If the team is new, name what is true: a shipping cadence so far, a domain expertise, a community track record.

1. Deliverables and milestones

• Three to five deliverables. Each has a definition of done that a reviewer can verify (a deployed package, a public repo, a documented integration).
• Milestones are dates with payouts. Each milestone bundles deliverables. Aim for three milestones for grants under 6 months, four to six for longer.

1. Budget

• Line items, dollar amounts. Engineering, design, audit, infra, contingency.
• Contingency capped at 10%. A higher contingency reads as poor planning.
• Tie line items to the deliverables: which deliverable consumes which line.

1. Public-good rationale

• Name the open-source license (MIT preferred unless there is a reason).
• Name the public artifacts (repo, docs, examples).
• Name the measurable benefit to the ecosystem (downstream projects will use, integrations will land, security findings will be shared).

1. Sustainability

• Honest answer. If the path is "we will charge for X starting M3", say so. If the path is "we will apply for follow-on grants", say so. If the path is "we will find a buyer or shut down", say so.
• "We will figure it out" fails this section. Reviewers read it as a flag.

1. Writeback and hand off

• Append .suiperpower/grant-application.md.
• Recommend roast-my-product to stress-test the application.
• Surface the grant portal URL when known.

Quality gate (anti-slop)

Before reporting done:

• Is every required input present, or were missing inputs surfaced and routed?
• Does each deliverable have a verifiable definition of done?
• Does each milestone tie a date to a payout to a deliverable?
• Is the budget line-itemed, with contingency under 10%?
• Does sustainability section give a real answer, not "we will figure it out"?
• Did the writeback happen?

If any answer is no, the skill keeps working.

References

On-demand references (load when relevant to the user's question):

• references/milestone-templates.md: Milestone shapes for 3, 6, and 12 month grants.
• references/budget-line-items.md: Common line items and pricing assumptions.

Use in your agent

• Claude Code: claude "/suiper:apply-grant <your message>"
• Codex: codex "/apply-grant <your message>"
• Grok Build: run grok, then /apply-grant <your message> in the session
• Cursor: paste a chat message that includes a phrase like "apply for a Sui Foundation grant", or load ~/.cursor/rules/apply-grant.mdc and reference it.

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.