Bottleneck Research CLI

Bottleneck Research CLI exports public Bottleneck Research data into agent-readable JSON or Markdown. Use it inside Codex, Claude Code, Cursor, notebooks, shell workflows, or your own research agents.

English | 中文

Website: bottleneckresearch.com
Public data feed: bottleneckresearch.com/data.json

What It Does

The CLI turns AI supply-chain bottleneck research into structured context:

• current research view and market regime
• AI infrastructure and application-layer candidate pools
• chain-level views for optical, storage, power, PCB/CCL, passive components, packaging and testing
• evidence freshness, missing proof and source quality checks
• decision-check context for a single ticker
• compact context blocks for external agents

It is designed for research workflows where a user asks an agent a question such as:

Can I buy 06088.HK?

The CLI does not answer with buy/sell instructions. It converts that question into an evidence, valuation, freshness, crowding and counter-evidence review.

Install

Recommended:

pipx install git+https://github.com/chatjesus/bottleneck-research-cli.git

Alternative:

uv tool install git+https://github.com/chatjesus/bottleneck-research-cli.git

Or install with pip:

python -m pip install git+https://github.com/chatjesus/bottleneck-research-cli.git

Run from a local clone:

git clone https://github.com/chatjesus/bottleneck-research-cli.git
cd bottleneck-research-cli
python br_research_cli.py context --format markdown

After installation, the command is:

br

Quick Start

The public feed is used by default. No API key is required.

br context --format markdown
br decision-check 06088.HK --format markdown
br candidates --chain optical --limit 10 --format markdown
br freshness --format markdown

If you want to specify the public endpoint explicitly:

br --base-url https://bottleneckresearch.com context --format markdown

For offline testing:

br --data-file tests/fixtures/sample_data.json context --format markdown

Common Commands

Agent Context

br agent-context --format markdown

Emits a compact research context that can be pasted into Codex, Claude Code, or another agent.

Decision Check

br decision-check 06088.HK --format markdown

Returns:

• research bucket
• evidence completeness
• missing orders, capacity, ASP, EPS/revenue expectation or management disclosure evidence
• price-volume status
• crowding risk
• freshness status
• next research action

Example output shape:

# Decision Check: 06088.HK

- research_bucket: `evidence_incomplete_continue_diligence`
- not_buy_sell_instruction: `true`
- evidence_score: 0.0
- price_volume_status: `unconfirmed`
- crowding_risk: `unknown`

## Next Research Action
Verify customer revenue split, orders, capacity, margin and management disclosure.

Chain View

br chain storage --format markdown
br candidates --chain optical --limit 10 --format markdown
br graph --chain power --format json

Freshness and Risk Checks

br freshness --format markdown
br macro --format markdown
br signals --format markdown

Compare Tickers

br compare 06088.HK 00894.HK 01888.HK --format markdown

Agent Usage Pattern

Run:

br decision-check 06088.HK --format markdown

Then ask your agent:

Based on this Bottleneck Research context, convert "can I buy 06088.HK?"
into an evidence, valuation, freshness, crowding and counter-evidence review.
Do not provide personalized trading advice.

Public Data and API Keys

The open-source CLI reads:

https://bottleneckresearch.com/data.json

No API key is required for the public endpoint.

The CLI supports --api-key and BR_API_KEY for future protected endpoints, but they are not needed for the current public feed.

Research Boundary

Bottleneck Research CLI is a research context tool. It helps organize public market data, supply-chain evidence, candidate pools, freshness checks and risk signals for further diligence.

It is not:

• investment advice
• a buy/sell recommendation engine
• a trading signal service
• a portfolio or position-sizing tool
• a substitute for independent diligence

All outputs should be reviewed against primary sources, valuation, liquidity, risk tolerance and personal suitability.

Development

git clone https://github.com/chatjesus/bottleneck-research-cli.git
cd bottleneck-research-cli
python -m unittest discover -s tests
python br_research_cli.py --data-file tests/fixtures/sample_data.json schema

Contributing

Issues and pull requests are welcome when they improve research context quality, data provenance, agent interoperability, documentation or test coverage. See CONTRIBUTING.md.

Security

Please do not open public issues for sensitive data exposure or security problems. See SECURITY.md.

License

MIT License. See LICENSE.