The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The script contains a hardcoded Twitter Bearer Token in scripts/cdp_tweet_fetcher.py and scripts/fetch_articles.py. While this is a known public token used by the Twitter web client, it is hardcoded as a static string in the source code.
[DATA_EXFILTRATION]: The skill performs sensitive data access by extracting auth_token and ct0 (CSRF) cookies from the user's browser via a Playwright CDP connection. These cookies are used to authenticate direct HTTP requests to Twitter's internal GraphQL endpoints.
[COMMAND_EXECUTION]: Instructions in SKILL.md require the user to execute shell commands to launch their browser with the --remote-debugging-port flag enabled, which allows the skill's scripts to connect to and control the active browser session and its data.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from an external platform. Ingestion points: Data enters the system via Twitter's GraphQL API responses in cdp_tweet_fetcher.py and fetch_articles.py. Boundary markers: None; the skill does not use delimiters or provide instructions to the agent to ignore embedded commands in the scraped content. Capability inventory: The skill has the capability to write files to the local disk and make outbound network requests via httpx. Sanitization: The skill converts Draft.js content to Markdown but does not perform sanitization or escaping of the raw tweet text before saving it to JSON or Markdown reports.
[EXTERNAL_DOWNLOADS]: The skill downloads and installs the playwright and httpx Python packages from standard registries and fetches the Chromium browser binary via Playwright. These are documented as well-known dependencies.

x-cdp-scraper