slack

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted text from a Slack workspace.\n
  • Ingestion points: The skill retrieves message content, channel topics, and user information from the browser DOM using agent-browser get text and agent-browser snapshot --json (found in SKILL.md and references/slack-tasks.md).\n
  • Boundary markers: Missing. No delimiters or 'ignore' instructions are used when interpolating Slack content into the agent's context.\n
  • Capability inventory: The skill utilizes Bash to run agent-browser, which has capabilities for browser navigation, data extraction, and writing files (screenshots/JSON) to the local environment.\n
  • Sanitization: None detected. The skill does not validate or sanitize Slack message content before processing.\n- [EXTERNAL_DOWNLOADS]: The skill uses the npx agent-browser command to execute its core functionality.\n
  • This command fetches and executes the agent-browser package from the NPM registry at runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 06:39 AM