Financial Times Deep Reader

Automates login to FT.com to extract and provide detailed bilingual English-Chinese summaries of top Financial Times articles with academic rigor.

Security Scan

The skill's stated purpose (logging into ft.com and summarizing articles) matches its runtime instructions, but it asks the agent to use a browser profile and credentials that are not declared or justified in the registry metadata, which is an important inconsistency and a potential credential exposure risk.

Purpose & Capability

The skill's purpose is to log into FT.com and extract articles — that legitimately requires credentials or session access. However, the registry lists no required environment variables, no primary credential, and no required config paths, while SKILL.md explicitly instructs use of a Browser profile named 'openclaw' and to enter email/password. This mismatch (claims of automated login vs. no declared credential/config requirements) is incoherent and raises risk about where credentials would come from and how they would be accessed.

Instruction Scope

SKILL.md instructs the agent to navigate to FT login pages, enter email/password, and run page-evaluation JavaScript to extract article text — these actions are consistent with the stated functionality. However, the instructions also demand use of a named browser profile ('openclaw') to maintain session persistence and include placeholder credentials in the file, without describing the source/secure storage of those credentials. That gives the skill broad discretion to access browser session data and credentials not declared to the registry, which is outside what a safe, self-contained instruction-only skill should assume. Also note potential legal/ToS concerns about automated scraping of paywalled content; the skill does not mention compliance.

Install Mechanism

There is no install spec and no code files — the skill is instruction-only. This minimizes filesystem risk because nothing is downloaded or written during install. The lack of an install step also explains why there are no declared dependencies, but it does not resolve the credential/access inconsistencies.

Credentials

The skill needs login credentials and persistent browser session access to function, yet the registry lists no required env vars, no primary credential, and no config paths. Requiring access to a browser profile (which can contain cookies, tokens, and other site credentials) is a high-privilege action; it should be explicitly declared and justified. The omission of any declared credential or config requirements is disproportionate and ambiguous.

Persistence & Privilege

The skill is not set to always:true and does not request persistent installation. Autonomous invocation is allowed (the platform default). Combined with the credential/session access the SKILL.md asks for, autonomous runs could access stored browser sessions or prompt for credentials without clear registry-level controls — this increases the blast radius but is not itself a configuration flag set by the skill.

What to consider before installing

This skill is suspicious because its runtime instructions require login credentials and use of a specific browser profile, but the registry metadata does not declare any credentials or configuration paths. Before installing or enabling it: 1) Ask the publisher for source code or a homepage and for a clear explanation of where credentials should be stored and how they are protected (prefer OAuth or token-based flows over plaintext credentials). 2) Do not enter your FT account email/password unless you fully trust the publisher; prefer using an account with no subscriptions or a throwaway account for testing. 3) Ask whether the skill will access your local browser profile or cookies; if so, decline unless you understand exactly what data will be read and why. 4) Confirm legal/ToS implications of scraping paywalled content. 5) If you must test, run it in a restricted environment (isolated profile or VM) and monitor network traffic. Providing a trusted source repo, explicit required env var names, or an OAuth-based integration would increase confidence and could change this assessment.

Financial Times Deep Reader (ft-reader)

Use this skill to perform deep, structured, and bilingual analysis of top articles from Financial Times (ft.com). This skill automates login, article selection, and high-quality summarization suitable for academic and professional use.

Capabilities

• Automated Access: Logs into FT.com using stored credentials via Browser tool.
• Strategic Selection: Identifies "Most Read" based on user preference.
• Bilingual Synthesis: Provides high-fidelity English-Chinese summaries with a focus on core arguments.
• Academic Rigor: Extracts specific data, quotes, and important charts in the article.

Configuration & Credentials

• Browser Profile: Use openclaw profile to maintain session persistence.
• Credentials:
  • User: xxxxxx
  • Pass: xxxxxx

Workflow (Mandatory Steps)

Phase 1: Authentication & Navigation

1. Open https://www.ft.com/login.
2. Enter email and password.
3. Navigate to the homepage or a specific section requested by the user.

Phase 2: Content Extraction

1. Use evaluate to identify the top N articles from the homepage (targeting .o-teaser__heading or most-read sections).

2. For each target article:

   • Navigate to the article URL.

   • Use evaluate with the following JavaScript to extract clean content:

     () => {
       const title = document.querySelector('h1')?.innerText;
       const standfirst = document.querySelector('div[class*="standfirst"]')?.innerText;
       const paragraphs = Array.from(document.querySelectorAll('div[class*="article-body"] p, article p'))
         .map(p => p.innerText.trim())
         .filter(text => text.length > 0);
       return { title, summary: standfirst, content: paragraphs.join('\n\n') };
     }
     

Phase 3: Analysis & Reporting

For each article, generate a report (around 600 words) using the following structure:

• Title (Bilingual)
• Core Opinion (Bilingual)
• Arguments (Bilingual)
• Conclusion (Bilingual)

Constraints

• Style: Professional, academic, and fluff-free (follow SOUL.md).
• Language: Always provide both English and Chinese translations for technical terms and core ideas.
• Independent Reading: Treat each article as a standalone piece unless cross-analysis is requested.
• Token Management: If many articles are requested, split the delivery into multiple turns to avoid truncation.

Usage Examples

• "Lulu, use ft-reader to analyze the top 3 Most Read articles from today."
• "Perform a deep dive into the top story on FT regarding AI productivity using the ft-reader skill."