openclaw-tradingview-quant
Fail
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill repeatedly directs the agent and users to
https://www.tradingviewapi.comfor documentation and API access (e.g., inREADME.md,references/api-documentation.md, andreferences/api-examples/README.md). This domain is not the official TradingView domain (tradingview.com) and has been flagged as malicious by automated security scanners. Interacting with this domain presents a critical risk of malware exposure and credential theft. - [PROMPT_INJECTION]: The skill processes untrusted third-party financial news and calendar data, which constitutes an indirect prompt injection attack surface.
- Ingestion points: External data enters the agent's context through the
tradingview_get_newsandtradingview_get_calendartools, specifically as documented inSKILL.mdand thenews-briefing.mdandevent-analysis.mdworkflows. - Boundary markers: The skill's
SKILL.mdfile contains a 'Security and Content Safety' section instructing the agent to treat news content as untrusted and ignore any embedded commands. - Capability inventory: The agent's capabilities include performing network API calls to fetch market data and generating detailed investment reports and strategy recommendations.
- Sanitization: The skill provides instructions for the agent to sanitize content by ignoring directive-like language and focusing strictly on factual market data.
Recommendations
- AI detected serious security threats
- Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata