# claude-equity-research: a Claude Code plugin that writes sell-side style equity reports > claude-equity-research is a Claude Code plugin that turns a ticker symbol into a structured equity research report with a rating, a price target, bull, base and bear scenarios and position sizing guidance. It is authored by carolinacherry, released under MIT, and it depends on live web data plus a paid Claude subscription rather than on any bundled market data. **quant-sentiment-ai/claude-equity-research** — 🔌 Claude Code Plugin for institutional-grade equity research. Install with /plugin marketplace add. Generates professional buy/sell recommendations with comprehensive fundamental analysis, technical indicators, and risk assessment. Educational use only - not financial advice. - Repository: https://github.com/quant-sentiment-ai/claude-equity-research - Stars: 742 · Forks: 90 - Language: Python - License: MIT - Published: 2026-10-21 · Updated: 2026-10-21 · Language: en - Canonical page: https://hysenlabs.com/projects/quant-sentiment-ai-claude-equity-research ## A packaged prompt that produces a fixed research report shape claude-equity-research is a Claude Code plugin, and the plugin's description says it generates buy and sell recommendations with fundamental analysis, technical indicators and risk assessment, with a note that it is for educational use only and is not financial advice. The interesting thing about the repository is how little code it needs. The primary language is recorded as Python, but the substance is a command file, research.md, that lives under commands/trading-ideas/commands/. That file is a specification for the analysis: what to retrieve, how to organise it, which sections to write and what disclaimers to attach. The intelligence comes from the model and from live web search, not from a computation library. That design has a clear consequence for expectations. There is no market data pipeline, no factor model, no backtest and no price history in this repository. The report is generated from what the model can find on the web at the moment it runs, so the same command run twice on the same ticker can differ. The feature list is ambitious in scope. It lists institutional-style output with a thesis, price target and twelve-month timeframe, fundamental analysis of revenue growth and margins with peer comparison, catalyst analysis with dates, bull, base and bear valuation scenarios with probability weighting, company-specific and macro risk with position sizing guidance of 1 to 5 percent, and technical context covering support and resistance, momentum and volume. An enhanced layer adds options flow, insider activity, sector positioning, and ESG and governance assessment. ## Installing through the plugin marketplace and verifying the command The prerequisites are short but not optional. You need the Claude Code CLI at version 2.0.11 or higher, a paid Claude subscription on the Pro, Team or Enterprise plan, and an internet connection, because the plugin retrieves live market data, earnings reports and analyst coverage at run time. The recommended path adds the repository as a marketplace and then installs from the plugin menu. ```bash /plugin marketplace add quant-sentiment-ai/claude-equity-research /plugin install trading-ideas@claude-equity-research-marketplace ``` After installing, you check that the command is registered. ```bash /help # Confirm /trading-ideas:research command is listed ``` The namespace detail catches people out. Claude Code namespaces plugin commands as plugin-name followed by command-name, so the invocation carries a colon and the bare form does not work unless you take the manual path. ```bash /trading-ideas:research AAPL /trading-ideas:research NVDA --detailed ``` The README also advises restarting Claude Code after installation for the best results, and points to PLUGIN.md in the repository for the fuller plugin documentation. Everything here happens inside a Claude Code session rather than in a terminal of your own, which is the reason there is no package to install, no dependency to pin and nothing to run. ## The manual path copies a single markdown file into your commands directory For people who want the bare command name instead of the namespaced one, the README documents a manual install that fetches the same research.md file directly into your personal commands directory. ```bash mkdir -p ~/.claude/commands curl -o ~/.claude/commands/trading-ideas.md https://raw.githubusercontent.com/quant-sentiment-ai/claude-equity-research/main/commands/trading-ideas/commands/research.md ``` The clone and copy route is the same thing with git in the middle. ```bash git clone https://github.com/quant-sentiment-ai/claude-equity-research.git cd claude-equity-research ``` ```bash cp commands/trading-ideas/commands/research.md ~/.claude/commands/trading-ideas.md ``` Either way you end up with a single markdown file in ~/.claude/commands/ and the command available without the :research suffix. This is worth understanding as a design choice rather than a shortcut. The plugin marketplace route gives you version tracking and updates, while the manual route gives you a file you can read, edit and version-control yourself. If you want to change the report structure, for example to drop the technical section or to add your own risk categories, editing that file is the whole workflow, because there is no application logic around it. The cost of the manual route is that you own the updates. The file is fetched from the main branch of the repository, so it changes under you, and there is no release history in the repository to pin against. ## What the report actually contains, section by section The README includes a sample output block, and its structure is the specification. It opens with an executive summary that carries the rating, the price target, an upside figure and a twelve-month timeframe, then a fundamental section with revenue, earnings per share and segment figures, then a valuation section with a consensus range and bull, base and bear cases, and closes with a recommendation line that pairs the rating with a conviction level and the price target. The bull, base and bear cases are given probability weights, which is the detail that distinguishes this from a prompt that just asks for an opinion. A probability-weighted scenario set gives the reader something to disagree with precisely, since the weighting is where a view is actually expressed. The command reference sets out three invocations. A ticker alone produces a standard institutional analysis in an eight-section report. A ticker with --detailed produces an enhanced analysis with options flow, extended technical and insider analysis. A --help flag shows the usage information for the command itself. The README gives one worked example of the sector-aware behaviour, describing the technology sector invocation as focused on AI and semiconductor sector positioning and relative valuation against peers, and another on financial services with interest rate sensitivity, regulatory environment and book value analysis. Those specialisations are what make it a research template rather than a generic summariser, and they are also the part that a user is most likely to want to edit. ## The repository is a plugin package, not an application The top-level file list explains the shape of the project: a .claude-plugin directory for the plugin manifest, a commands directory holding the command definitions, a config directory, a docs directory, an examples directory, a scripts directory, plus PLUGIN.md, SECURITY.md, LICENSE and the README. The examples directory holds sample reports under examples/sample_reports/. Two of those entries matter for anyone considering the plugin seriously. PLUGIN.md is the detailed plugin documentation the README points to, and SECURITY.md is a file that a plugin distributed through a marketplace ought to have, since it is where the disclosure of what the plugin does with your data would live. The absence of a package manifest is the other signal. There is no pyproject.toml, no requirements file and no build step, which is consistent with the project being a set of prompts and configuration rather than a Python program. The recorded primary language reflects the small amount of supporting script code in the scripts directory, not a runtime that a user starts. The project is by carolinacherry, and the README says it was built entirely with Claude Code, describing that as a demonstration of AI-native development workflows. The badge in the README marks the status as Active, and the repository has no GitHub releases, so the versioning story is the plugin marketplace rather than tagged archives. ## Running the plugin costs a subscription and a web connection, not a licence fee The licence is MIT, with the LICENSE file at the repository root, so the code and the prompts can be reused and modified without restriction beyond keeping the notice. There is no separate fee for the plugin itself. The real cost of running it is elsewhere. It needs a paid Claude subscription, and it performs live web retrieval on every invocation, so the token and search cost scales with how often you run it and how deep the analysis goes. The --detailed mode is the expensive one, since the enhanced layer adds options flow and insider analysis on top of the standard report. On the maintenance side, the repository is not archived, the last push was on 2026-10-17, and there are no GitHub releases. What that means in practice is that the plugin marketplace entry is the release channel, while a manual install pulls the file from the main branch and therefore tracks whatever the newest content is rather than a versioned artefact. If you rely on the report format, expect it to change under you when you installed manually. The report includes a legal protection section with disclaimers for educational use, which the README lists as a professional standard alongside the institutional terminology and the analyst firm citations from real-time web search. That disclaimer is not decoration: it is the project's own statement of what the output is and is not. ## The plugin against a hand-written prompt in your own session There is a real alternative to installing this, and it is the approach many readers will already be using: writing your own analysis prompt in a Claude Code session and iterating on it. The difference in approach is about shape rather than capability. A hand-written prompt produces whatever structure you asked for on the day you wrote it, and you can change it between runs, but there is no shared report format, no consistent section list and no agreed vocabulary between runs, so two reports are hard to compare. The plugin fixes an eight-section structure, a standard set of institutional terms, probability weighting for the scenarios, a position sizing range and a disclaimer block, which makes two reports for two tickers comparable and makes a reviewer able to check the same fields in the same order every time. The trade-off runs the other way too. Because the structure is fixed in a file you may not have read, the plugin can also push a report into a shape that does not fit your question, and when the framing is wrong the fix is to edit the command file rather than to re-prompt. Teams that need a bespoke report for a specific mandate are better served by their own prompt; teams that need many tickers examined the same way are better served by the packaged command. For a quick evaluation, the manual install is the lower-commitment route, since it is one curl into a directory you can delete. ## Conclusion claude-equity-research fits a developer or analyst who wants a repeatable, well-shaped research document for a ticker in seconds rather than a price feed or a terminal, and it fits teams that treat the output as a drafting aid. It does not fit anyone who needs verified prices, backtested signals or a compliance-grade recommendation, since the reports are generated prose drawn from live web search and the README states the tool is for educational use and is not financial advice. Before you install it, read the disclaimers in the report template and the SECURITY.md file, check that a paid Claude plan and a Claude Code CLI of version 2.0.11 or newer are available to you, and treat any price target it prints as a model output rather than a quote. ## FAQ ### How do I install the claude-equity-research plugin? Add the repository as a marketplace with `/plugin marketplace add quant-sentiment-ai/claude-equity-research` and install with `/plugin install trading-ideas@claude-equity-research-marketplace`, or fetch the research.md command file manually into ~/.claude/commands/ to use the bare `/trading-ideas` form. The README recommends restarting Claude Code after installing. ### What do I need before I can run the equity research command? The README lists the Claude Code CLI at version 2.0.11 or higher, a paid Claude subscription on the Pro, Team or Enterprise plan, and an internet connection for real-time data retrieval. ### What does the trading-ideas research command output? A ticker alone returns an eight-section institutional-style report with a BUY, SELL or HOLD recommendation, a price target, probability-weighted bull, base and bear scenarios and position sizing guidance of 1 to 5 percent. Adding --detailed extends the analysis with options flow, technical and insider detail, and --help shows the command usage. ### Is the output from claude-equity-research financial advice? No. The repository description states the tool is for educational use only and is not financial advice, and the README lists disclaimers for educational use as one of its professional standards. The reports are generated by the model from live web sources rather than from bundled market data. ### What licence does claude-equity-research use? The project is MIT licensed, with the licence file at the repository root. It has no GitHub releases, so the plugin marketplace and the manual file copy from the main branch are the two distribution paths. ## Sources - [Issues](https://github.com/quant-sentiment-ai/claude-equity-research/issues) - [License: MIT](https://github.com/quant-sentiment-ai/claude-equity-research/blob/main/LICENSE) - [quant-sentiment-ai/claude-equity-research on GitHub](https://github.com/quant-sentiment-ai/claude-equity-research) - [README](https://github.com/quant-sentiment-ai/claude-equity-research/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/quant-sentiment-ai-claude-equity-research