baoyu-danger-x-to-markdown
Warn
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The file
scripts/constants.tscontains a hardcoded bearer token (DEFAULT_BEARER_TOKEN) used for authenticating with the X API. - [COMMAND_EXECUTION]: The script
scripts/cookies.tsuseschild_process.spawnto launch browser executables (Chrome, Chromium, or Edge) with the--remote-debugging-portflag enabled. This is used to programmatically intercept authentication cookies. - [EXTERNAL_DOWNLOADS]: The skill fetches data from
x.comand downloads media assets fromtwimg.comdomains. These are well-known services associated with the skill's primary function. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing untrusted data from X (Twitter) posts and articles.
- Ingestion points: Data is ingested through API calls in
scripts/graphql.tsandscripts/thread.tswhich retrieve content from external URLs. - Boundary markers: None. The skill does not add delimiters or explicit warnings to the output markdown to prevent an agent from interpreting the ingested content as instructions.
- Capability inventory: The skill environment includes file system writing (
writeFile), network access (fetch), and subprocess execution (spawn). - Sanitization: Basic sanitization is applied to filenames and URLs, but the body content of tweets and articles is not filtered for potential malicious instructions.
Audit Metadata