skills/azmym/agent-skills/slack/Gen Agent Trust Hub

slack

Fail

Audited by Gen Agent Trust Hub on Mar 26, 2026

Risk Level: HIGHCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The slack-token-refresh.sh script extracts authentication cookies (d cookie) and session tokens (xoxc) from the local Google Chrome installation on macOS.
  • It locates and copies the Chrome Cookies database to a temporary directory.
  • It uses AppleScript to extract Slack's localStorage tokens from active browser tabs.
  • [COMMAND_EXECUTION]: The skill uses high-privilege system commands and scripting interfaces to harvest credentials.
  • Uses security find-generic-password -s 'Chrome Safe Storage' -w to programmatically access the macOS Keychain and retrieve the decryption key for the browser's cookie database.
  • Uses osascript (AppleScript) to execute JavaScript inside running Google Chrome tabs, which can bypass some browser security boundaries.
  • [EXTERNAL_DOWNLOADS]: The skill downloads and executes code from remote repositories at runtime.
  • The slack-token-refresh.sh script uses uvx --from pycookiecheat to download and run the pycookiecheat package from PyPI.
  • The slack-browser-session.sh script performs an unversioned npm install and npx playwright install chromium to download the Playwright framework and the Chromium browser engine.
  • [REMOTE_CODE_EXECUTION]: The playwright-bridge.js script contains an execute function that allows the execution of arbitrary JavaScript code within the browser context using page.evaluate(input.code). This provides a direct path for dynamic code execution of instructions supplied at runtime.
  • [DATA_EXFILTRATION]: Extracted authentication secrets (SLACK_XOXC and SLACK_XOXD) are stored in a local environment file at ~/.agents/config/slack/tokens.env. While local, this makes these sensitive credentials accessible to any process with read access to the user's home directory.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 26, 2026, 08:18 PM