Dossier DOSSIER SYSTEM // V2.0
Run Dossier
Zero-Sycophancy Idea Intelligence

Interrogate your venture before the market kills it.

Six domain-specialized agents interrogate your business model, stress-test unit economics with 10,000 Monte Carlo runs, and cross-reference real-world merchant demand with startup failure forensics.

INPUT CONSOLE // PLAIN TEXT SPECIFICATION
0 / 2,000
Sample ideas:
Design Philosophy

Why traditional LLMs hallucinate viable startups.

Standard frontier models are RLHF-tuned for agreeable sycophancy. When you ask them about an idea, they flatter your assumptions instead of stress-testing fatal operational hazards.

Standard Frontier LLM
Sycophantic Validation

Praises every feature request, assumes linear market adoption, and quotes generic TAM numbers without verifying procurement or statutory friction.

Static Financial Models
Deterministic Illusions

Spreadsheets calculate uniform payback periods using static assumptions, ignoring tail churn events, CAC inflation, and insolvent downside distributions.

Dossier Adversarial Swarm
Empirical Invariance & Proof

Deterministic Python sandbox Monte Carlo, Razorpay merchant demand grounding, Loot Drop failure traps, and Epistemic DAG verification with anti-sycophancy caps.

Agent Specifications

Six specialized adversarial agents.

Each agent operates under a strict mathematical and operational charter, probing distinct failure vectors before reaching final consensus.

AGENT 00 SYNTHESIZER
Executive Synthesizer

Establishes domain priors, identifies procurement cycles, and synthesizes crossfire arguments into actionable strategic verdicts.

AGENT 01 CONTRARIAN
The Skeptic (Red Team)

Probes existential distribution risks, enterprise sales deadlocks, incumbent bundling traps, and uncompensated tail risks.

AGENT 02 ENGINEERING
Technical Architect

Analyzes latency, workflow integration, model inference costs, API dependency moats, and data residency constraints.

AGENT 03 ORACLE & RECON
Market Recon

Queries Polymarket betting volumes, live news headlines, developer traction signals, and competitor displacement friction.

AGENT 04 REALITY CHECK
Customer Advocate

Matches propositions against real merchant problem catalogs (Razorpay Fix My Itch) and validates daily workflow switching friction.

AGENT 05 VENTURE CAPITAL
Financial Analyst

Calibrates CAC payback bounds, gross margin thresholds, and verifies scores against isolated Python Monte Carlo output.

[+] DEVELOPER & MCP PROTOCOL INTEGRATION

Connect Dossier Adversarial Swarm to Claude Desktop, Cursor, & CLI

Expose the 6-agent swarm, isolated Python Monte Carlo runner, and epistemic verification engine over standard MCP stdio JSON-RPC.

View Stdio JSON Config 
{
  "mcpServers": {
    "dossier": {
      "command": "node",
      "args": ["/home/sanjeev/projects/products/dossier/dist/mcp-server.js"]
    }
  }
}
evaluate_idea — Full 6-agent swarm intelligence & dossier synthesis
run_sandbox_economics — Isolated Python3 deterministic unit economics
verify_claims_epistemics — 7-tier epistemic hierarchy claim audit
Audit Trail

Recent Verified Dossiers

Loading audit archive...