The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file scripts/constants.ts contains a hardcoded DEFAULT_BEARER_TOKEN. While this is a known token for the X web client, hardcoding credentials in the skill source is an unsafe practice.
[COMMAND_EXECUTION]: The skill executes external commands for essential functionality, which increases the potential attack surface.
Evidence: scripts/cookies.ts uses spawn to launch Google Chrome for the purpose of automated cookie extraction.
Evidence: scripts/paths.ts uses execSync to run cmd.exe for resolving Windows home paths in WSL environments.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from an external social platform.
Ingestion points: Untrusted content is fetched from X in scripts/graphql.ts and scripts/thread.ts.
Boundary markers: Content is formatted into markdown without specific delimiters or instructions for the agent to ignore embedded commands.
Capability inventory: The skill has the ability to write files (scripts/main.ts) and execute shell commands (scripts/cookies.ts).
Sanitization: There is no logic present to sanitize or escape potentially malicious instructions embedded in the tweet or article content.

baoyu-danger-x-to-markdown