Biz Analysis Pro — Business Competition Pipeline (商赛流水线)

    1

    The pipeline enforces discipline; you drive judgment.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Biz Analysis Pro — Business Competition Pipeline (商赛流水线)

    Example session with this skill installed

    We're entering a business case competition about the specialty coffee market in China. Here is the problem statement PDF. Produce the full analysis: industry scan, market sizing, competitive landscape, strategy and a 3-year financial model.

    • Read your context and instructions
    • Compiled the biz analysis pro

    work_dir/ with: industry_scan.md (every number cited [source, year]), market_sizing.md (TAM/SAM/SOM × conservative/base/optimistic), competitive_analysis.md (Porter + competitor matrix + positioning chart), customer_analysis.md, strategy.md (BMC + GTM), financial_model.py outputs (yearly projection, LTV/CAC, break-even), risk matrix, figures/ (7 chart types), report/report.md following the 9-chapter template.

    industry_scan.md

    Connects securely to your tools. The creator never sees your data.

    About this skill

    What this is

    A 12-stage pipeline for business analysis competitions and case studies, run as specialist subagents (Scout / Analyst / Strategist / Financier / Writer / Reviewer) with ready-to-paste prompts for every stage.

    Honest positioning: this is interactive, not autopilot. There are 6 mandatory checkpoints where the pipeline stops and you decide: problem framing, research scope, strategy direction, financial assumptions, draft review, final sign-off. The pipeline enforces discipline; you drive judgment. If you reply "随便/你定/default", it proceeds with its recommendation instead of re-asking.

    The pipeline

    Problem intake → industry scan (web research) → market sizing (TAM/SAM/SOM) → competitive landscape (Porter + matrix + positioning) → customer analysis → strategy design (BMC + GTM) → financial model (3-5yr projections) → risk assessment → report writing → slides (optional) → self-review against scoring rubric → final output.

    Hard data-discipline gates (the core value)

    • Every market-size claim must cite a source (report name + year, or shown calculation). No fabricated industry data.
    • Every financial projection must list assumptions explicitly in a table before any numbers.
    • Common-mistake table included: "市场规模很大" without a number, Porter template without industry-specific evidence, financials without assumptions — each with its fix.

    What's in the package

    • references/market-sizing-methods.md — 3 methods (top-down / bottom-up / value theory) with worked examples, cross-validation rule (2 methods, gap >50% → recheck), conservative/base/optimistic ranges, and where to find each parameter.
    • references/biz-charts.md — 7 copy-paste matplotlib templates (market waterfall, positioning map, revenue projection, funnel, risk matrix, Porter five forces, business model canvas) with save checklist.
    • references/report-template.md — full 9-chapter report skeleton with writing rules (executive summary must contain numbers, figure/table citation format).
    • references/scrapling-collect.md — 4 deep-collection templates for when WebSearch isn't enough: batch competitor sites, PDF industry reports, anti-bot/Cloudflare pages, financial-table extraction — with an escalation ladder (get → fetch → stealthy-fetch) and source-annotation rules that satisfy the hard gates.
    • scripts/financial_model.py — ready-to-run CLI: monthly user growth/churn/ARPU → yearly projection table, LTV, CAC, LTV/CAC, break-even month; markdown or JSON output.

    Requirements

    Works best in agents with web search + subagent dispatch (e.g. Claude Code); stages can also run sequentially in a single thread. Charts need Python + matplotlib. Scrapling and markitdown are optional upgrades, not dependencies.

    Not autonomous by design: 6 mandatory user checkpoints — you make the scope/strategy/assumption calls (replying "default/随便" lets it proceed with its recommendation). Subagent dispatch works best in Claude Code; in simpler agents, run the stages sequentially in one thread. Charts need Python + matplotlib. Scrapling and markitdown are optional, not required.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 9 days ago

    • Free with an Agensi account

    Requires

    Python 3.8+
    Matplotlib
    Python

    Permissions

    Terminal / Shell
    Write Files
    Browser
    Network Access
    Listed9 days ago

    What's inside

    Frequently Asked Questions

    Deliberately not. It scaffolds research → strategy → financials and stops at 6 checkpoints for your decisions. What IS enforced automatically: every market number cited, every assumption explicit, every recommendation traceable to a finding.

    Yes. The Scrapling reference becomes inactive; the pipeline runs on WebSearch + reasoning. Scrapling only unlocks deeper collection: PDF tables, batch competitor pages, anti-bot sites.

    Chinese-market sources are pre-mapped (艾瑞/易观/前瞻/国家统计局/annual reports), but the pipeline, frameworks and financial script are market-agnostic.

    New in Business & Operations