Grant Warfare Terminal

AI-powered tactical intelligence for DAO grant proposals.

Analyzes grant proposal text (psychology + competitive edge), caches results in Supabase, and surfaces a tactical brief with opportunity score, strategy, and export.

Live

GitHub: https://github.com/intenxe-ops/grant-terminal

System architecture

High-level

┌─────────────────────────────────────────────────────────────────────┐
│                         GRANT WARFARE TERMINAL                        │
│                     Next.js 16 · App Router · React 19               │
└─────────────────────────────────────────────────────────────────────┘
                                   │
                    ┌──────────────┴──────────────┐
                    ▼                             ▼
        ┌───────────────────────┐    ┌───────────────────────┐
        │   FRONTEND             │    │   API ROUTES           │
        ├───────────────────────┤    ├───────────────────────┤
        │ • ProposalScanner      │    │ POST /api/analyze      │
        │ • TacticalBriefPanel   │    │ POST /api/mcp          │
        │ • FilterBar            │    │ (same pipeline)        │
        │ • GlyphNavigation      │    │                         │
        │ • CommandHeader        │    │ Cache check → agents    │
        └───────────┬────────────┘    └───────────┬────────────┘
                    │                             │
                    │    POST body:                │
                    │    proposalId, title,       │
                    │    description, dao         │
                    └─────────────┬───────────────┘
                                  ▼
        ┌───────────────────────────────────────────────────┐
        │              lib/supabase.ts                        │
        │  getCachedAnalysis(proposalId)                     │
        │  Table: proposal_analyses                          │
        │  Hit → return cached. Miss → continue.              │
        └───────────────┬───────────────────────────────────┘
                        │ cache miss
                        ▼
        ┌───────────────────────────────────────────────────┐
        │              lib/agents/                            │
        ├───────────────────────────────────────────────────┤
        │  runPsychologyAnalysis({ title, description, dao }) │
        │  → Anthropic Messages API                           │
        │  → urgency, desperation, flexibility, triggers,    │
        │    fullAnalysis                                     │
        └───────────────┬───────────────────────────────────┘
                        │
                        ▼
        ┌───────────────────────────────────────────────────┐
        │  runEdgeAnalysis({ title, description,             │
        │    psychologyAnalysis })                            │
        │  → Anthropic Messages API                           │
        │  → skillMatch, winProbability, advantages, gaps,     │
        │    mainStrategy, emphasize, avoid, openingLine      │
        └───────────────┬───────────────────────────────────┘
                        │
                        ▼
        ┌───────────────────────────────────────────────────┐
        │  Opportunity score:                                 │
        │  (skillMatch × 0.4 + winProbability × 0.6) / 10    │
        │  saveAnalysis(record) → proposal_analyses           │
        └───────────────────────────────────────────────────┘

Data flow: analysis execution

USER: Click "ANALYZE TARGET" on a proposal
       │
       ▼
Frontend (ProposalScanner) POST /api/analyze
  body: { proposalId, title, description, dao }
       │
       ▼
1. getCachedAnalysis(proposalId) → Supabase proposal_analyses
   • Found → return { success, cached: true, data }
   • Not found → continue
       │
       ▼
2. runPsychologyAnalysis({ title, description, dao })
   • Anthropic Messages API (model from ANTHROPIC_MODEL or default)
   • Prompt: grant psychology analyst, JSON schema for urgency, desperation,
     flexibility, triggers, fullAnalysis
   • Returns: PsychologyResult
       │
       ▼
3. runEdgeAnalysis({ title, description, psychologyAnalysis })
   • Anthropic Messages API
   • Prompt: grant tactical analyst, psychology summary as context
   • Returns: EdgeResult (skillMatch, winProbability, advantages, gaps,
     mainStrategy, emphasize, avoid, openingLine)
       │
       ▼
4. opportunityScore = round((skillMatch×0.4 + winProbability×0.6) / 10)
       │
       ▼
5. saveAnalysis(record) → Supabase proposal_analyses insert
       │
       ▼
6. Response: { success, cached: false, data: analysisData }
   Frontend shows TacticalBriefPanel with brief; export via export-brief.

Intelligence layer (lib/agents)

Two sequential steps; both use Anthropic SDK, no external agent framework.

Psychology (`lib/agents/psychology.ts`)

Input: title, description, dao
Output: PsychologyResult
- urgency, urgencyReason
- desperation, desperationReason
- flexibility, flexibilityReason
- triggers: string[]
- fullAnalysis: string
Model: ANTHROPIC_MODEL env or claude-sonnet-4-20250514
Implementation: Single Messages API call; response parsed as JSON and normalized.

Edge (`lib/agents/edge.ts`)

Input: title, description, psychologyAnalysis (PsychologyResult)
Output: EdgeResult
- skillMatch, winProbability (0–100, clamped)
- advantages: string[], gaps: string[]
- mainStrategy, emphasize: string[], avoid: string[], openingLine
Model: Same as psychology.
Implementation: Psychology summary passed in prompt; single Messages API call; JSON parsed and normalized.

Opportunity score

Computed in API route:

(edge.skillMatch * 0.4 + edge.winProbability * 0.6) / 10, rounded. Displayed as 0–10 in the brief.

API endpoints

POST /api/analyze

Used by: ProposalScanner (fetchWithRetry from lib/api-client.ts).
Body: { proposalId, title, description?, dao }
Response: { success, cached?, data?: TacticalBrief, error?, details? }
Flow: Cache → psychology agent → edge agent → score → save → return.

POST /api/mcp

Same pipeline as /api/analyze; response includes source: 'mcp'.
Body: Same.
Used by: Clients that call this route (e.g. lib/api-client-mcp.ts).

Data layer (Supabase)

Table: proposal_analyses
Usage:
- getCachedAnalysis(proposalId): select by proposal_id, order by created_at desc, maybeSingle().
- saveAnalysis(record): insert full analysis (proposal_id, proposal_title, dao, opportunity_score, psychology, edge, recommendation).
Types: AnalysisRecord in lib/supabase.ts (matches psychology + edge + recommendation shape).

No other tables or caches are used in the codebase.

Frontend

Proposals: Loaded via lib/snapshot.ts — fetchActiveProposals(filter) calls Snapshot Hub GraphQL; filter by DAO type (all / defi / gaming / other).
Analysis: ProposalScanner calls /api/analyze with proposal id, title, body, dao; shows loading then TacticalBriefPanel with TacticalBrief.
Brief: Opportunity score, psychology (urgency, desperation, flexibility, triggers, fullAnalysis), edge (skillMatch, winProbability, advantages, gaps), recommendation (mainStrategy, emphasize, avoid, openingLine).
Export: lib/export-brief.ts — markdown or JSON download (exportBriefAsMarkdown, exportBriefAsJSON, downloadFile, exportTacticalBrief).
UI: Tailwind, Framer Motion, Lucide, Radix (tooltip, slot); glyphs (Triangle, Diamond, Hexagon, Pentagon); CommandHeader, GlyphNavigation, FilterBar, ProposalCard, TacticalBriefPanel, TacticalLogo, TacticalNavTrigger, ErrorBoundary, SidebarContext, HomeLanding.

Tech stack

Layer	Stack
Runtime	Next.js 16 (App Router), Node, React 19
LLM	@anthropic-ai/sdk (Claude); model via ANTHROPIC_MODEL or default
Data	Supabase (PostgreSQL); table `proposal_analyses`
Proposals	Snapshot Hub GraphQL (lib/snapshot.ts)
UI	Tailwind CSS, Framer Motion, Lucide React, Radix UI (tooltip, slot), class-variance-authority, clsx, tailwind-merge, tailwindcss-animate
Validation	Zod (used where applicable)
Package mgr	pnpm 10; packageManager in package.json
Dev	TypeScript 5.2, ESLint, eslint-config-next, baseline-browser-mapping (devDep)

Project structure

GrantTerminal/
├── app/
│   ├── api/
│   │   ├── analyze/
│   │   │   └── route.ts          # POST; cache → lib/agents → save
│   │   └── mcp/
│   │       └── route.ts           # POST; same pipeline, source: mcp
│   ├── layout.tsx
│   ├── page.tsx
│   └── globals.css
├── components/
│   ├── scanner/
│   │   ├── ProposalScanner.tsx    # Load proposals, analyze, show brief
│   │   ├── ProposalCard.tsx
│   │   ├── FilterBar.tsx
│   │   └── ScannerStates.tsx      # EmptyState, LoadingState, ErrorState
│   ├── tactical-brief/
│   │   └── TacticalBriefPanel.tsx
│   ├── glyphs/
│   │   ├── TriangleGlyph.tsx
│   │   ├── DiamondGlyph.tsx
│   │   ├── HexagonGlyph.tsx
│   │   └── PentagonGlyph.tsx
│   ├── ui/
│   │   └── tooltip.tsx
│   ├── CommandHeader.tsx
│   ├── GlyphNavigation.tsx
│   ├── TacticalLogo.tsx
│   ├── TacticalNavTrigger.tsx
│   ├── ErrorBoundary.tsx
│   ├── SidebarContext.tsx
│   └── HomeLanding.tsx
├── lib/
│   ├── agents/
│   │   ├── psychology.ts          # runPsychologyAnalysis
│   │   ├── edge.ts                # runEdgeAnalysis
│   │   └── index.ts
│   ├── supabase.ts                # client, getCachedAnalysis, saveAnalysis, AnalysisRecord
│   ├── api-client.ts              # fetchWithRetry, APIError
│   ├── api-client-mcp.ts          # calls /api/mcp
│   ├── snapshot.ts                # fetchActiveProposals, formatTimeRemaining, etc.
│   ├── export-brief.ts            # markdown/JSON export
│   ├── utils.ts
│   └── snapshot.ts                # Snapshot Hub fetch, formatTimeRemaining, etc.
├── types/
│   ├── tactical-brief.ts          # TacticalBrief, PanelState
│   └── snapshot.ts                # Proposal, SnapshotProposal, DAOFilter, etc.
├── next.config.js
├── package.json
├── pnpm-lock.yaml
├── LICENSE
└── README.md

Installation and local run

git clone https://github.com/intenxe-ops/grant-terminal.git
cd grant-terminal

pnpm install

cp .env.example .env
# Set:
#   ANTHROPIC_API_KEY=
#   NEXT_PUBLIC_SUPABASE_URL=
#   NEXT_PUBLIC_SUPABASE_ANON_KEY=

pnpm dev
# Open http://localhost:3000

Scripts: dev, build, start, lint, typecheck (see package.json).

Environment variables

Variable	Required	Description
ANTHROPIC_API_KEY	Yes	Anthropic API key for lib/agents
NEXT_PUBLIC_SUPABASE_URL	Yes	Supabase project URL
NEXT_PUBLIC_SUPABASE_ANON_KEY	Yes	Supabase anon key
ANTHROPIC_MODEL	No	Model id (default: claude-sonnet-4-20250514)

Usage

Load proposals — App fetches active proposals from Snapshot (FilterBar: all / defi / gaming / other).
Filter — Switch DAO filter; proposals reload.
Analyze — Click ANALYZE TARGET on a proposal; frontend POSTs to /api/analyze; loading state; on success TacticalBriefPanel shows the brief.
Brief — Opportunity score (0–10), psychology block (urgency, desperation, flexibility, triggers, full analysis), edge block (skill match %, win probability %, advantages, gaps), recommendation (main strategy, emphasize, avoid, opening line).
Export — Export brief as markdown or JSON (export-brief.ts).

Performance notes

Analysis: Two sequential LLM calls (psychology then edge); typical total in the tens of seconds depending on model and network.
Cache: Repeat analyses for same proposal_id served from Supabase; no second LLM run.
Timeout: api-client uses 90s timeout and retries (fetchWithRetry).

License

MIT. See LICENSE.

Author

intenxe-ops

Name		Name	Last commit message	Last commit date
Latest commit History 18 Commits
app		app
components		components
lib		lib
supabase/migrations		supabase/migrations
types		types
.eslintrc.json		.eslintrc.json
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
components.json		components.json
next.config.js		next.config.js
package-lock.json		package-lock.json
package.json		package.json
pnpm-lock.yaml		pnpm-lock.yaml
postcss.config.js		postcss.config.js
tailwind.config.ts		tailwind.config.ts
tsconfig.json		tsconfig.json

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Grant Warfare Terminal

Live

System architecture

High-level

Data flow: analysis execution

Intelligence layer (lib/agents)

Psychology (`lib/agents/psychology.ts`)

Edge (`lib/agents/edge.ts`)

Opportunity score

API endpoints

POST /api/analyze

POST /api/mcp

Data layer (Supabase)

Frontend

Tech stack

Project structure

Installation and local run

Environment variables

Usage

Performance notes

License

Author

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

Grant Warfare Terminal

Live

System architecture

High-level

Data flow: analysis execution

Intelligence layer (lib/agents)

Psychology (lib/agents/psychology.ts)

Edge (lib/agents/edge.ts)

Opportunity score

API endpoints

POST /api/analyze

POST /api/mcp

Data layer (Supabase)

Frontend

Tech stack

Project structure

Installation and local run

Environment variables

Usage

Performance notes

License

Author

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Psychology (`lib/agents/psychology.ts`)

Edge (`lib/agents/edge.ts`)

Packages