Security audit
Security checks for vulnerabilities and agentic risk
This stock-analysis skill is mostly purpose-aligned, but users should review it because it sends stock and market queries to TTfox and makes inconsistent claims about whether the data is live or delayed.
Install only if you are comfortable with stock symbols, sector names, market parameters, and related investment interests being sent to TTfox. Treat all outputs as research, not trading advice, and assume the data may be delayed despite some live/real-time wording in the documentation.
The declared description understates the actual behavior: the skill is effectively a thin client to an external service and supports broader market retrieval functions than the title/description suggest. This mismatch can mislead users and operators about external data sharing, trust boundaries, and the real execution path, which is especially risky in a finance-related skill that may be broadly triggered.
The README makes a strong trust-building claim that analysis is based on 'real, live market data' and later discloses that the data is actually delayed, post-market, and non-real-time. In a finance-oriented agent skill, this inconsistency can mislead users and upstream agents into making time-sensitive decisions under false assumptions about freshness and reliability of the data.
Broad natural-language triggers like 'How is the market?' or 'Analyze AAPL' can cause accidental invocation during ordinary conversation, especially in agent platforms that auto-route by pattern matching. In this skill's context, unintended activation may trigger external requests to the backend intelligence server and present financial analysis when the user did not explicitly intend to use this tool.
The skill advertises and operationally requires outbound access to an external intelligence server, but it does not declare any tool scope or permissions boundary. This weakens user/admin visibility into what the skill can access and increases the risk of unintended data egress or unauthorized network use when the skill is auto-invoked.
Broad trigger phrases such as market-related questions and buying-decision prompts can cause the skill to activate during ordinary conversation, leading to unexpected external API calls or unrequested financial guidance. In this context, overbroad activation is more dangerous because the skill is tied to network access and finance advice, where accidental invocation can expose user intent and create compliance or trust issues.
The Chinese trigger set contains highly generic phrases like market, stock, and analysis terms that are likely to appear in normal conversation and can unintentionally activate the skill. Because activation may result in outbound requests and finance-oriented responses, ambiguous multilingual triggers expand the accidental-activation surface and make the behavior harder for users to predict.
The heading and instructions state this is the "English Edition" and direct the agent to serve "English-speaking users," which imposes a language constraint. There is no explicit user choice or opt-in mechanism described, so this can violate language/locale policy expectations.
The class description labels this as an English-only edition, and the code forces lang: en in request parameters. This imposes a specific language/locale choice without offering the user an option to select another language, which matches the policy's language/locale violation criterion.
The skill transmits user-supplied stock identifiers to an external third-party service (master.ttfox.com) without any visible user consent, disclosure, or data-minimization controls. In an agent context, external transmission can leak user interests, portfolio/watchlist information, or sensitive workflow context to a remote service, and the risk is amplified because the skill is built around repeated network calls to a non-local backend.
try:
url = f"{self.base_url}/api/analyze"
payload = {"stock_code": stock_code, "lang": "en"}
res = requests.post(url, json=payload, headers=self.headers, timeout=10)
if res.status_code == 200:
result = res.json()
if result.get("status") == "success":
The file states the system is powered by '5 legendary investment master models,' and the capability table also says '5-Master Stock Diagnosis.' However, the integration examples later say stock analysis triggers '11-Master deep diagnosis,' which contradicts the documented scope of the analysis engine.
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.
## Legal Disclaimer
1. **NOT Investment Advice**: This project and all its content (including but not limited to analysis reports, master scores, and suggestions) are provided for **Educational and Academic** purposes only.
2. **No Liability for Latency**: The system generates conclusions from **Delayed Data**. The conclusions may differ significantly from current market conditions. The authors and contributors shall NOT be held liable for any financial losses caused by data lag, analysis errors, or system failure.
3. **Risk Assumption**: Investing in stock markets involves extreme risk. Users assume full responsibility for any investment outcomes based on the information provided by this tool.
No suspicious patterns detected.