daily-news
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill modifies the agent's environment configuration files (~/.claude/CLAUDE.md and ~/.claude.json) to store workspace paths and configure a Browser MCP server.
- [EXTERNAL_DOWNLOADS]: The skill requires installation of third-party dependencies, including Python packages (pyyaml, feedparser, requests, beautifulsoup4) and Node.js packages (@browsermcp/mcp, agent-browser).
- [DATA_EXFILTRATION]: The skill uses Git and GitHub CLI tools to push generated content to external repositories as part of the website deployment feature.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests content from external websites and RSS feeds. This untrusted data is processed using prompts in references/prompts/summary.md without sufficient sanitization or boundary markers.
- Ingestion points: News content fetched via RSS or Browser MCP.
- Boundary markers: Absent in the summarization prompt.
- Capability inventory: The agent can execute Python, modify local config files, and use Git/GitHub CLI.
- Sanitization: No sanitization of external content is performed.
Audit Metadata