Security audit
Security checks for vulnerabilities and agentic risk
This skill appears purpose-built for Trump/Iran market sentiment monitoring, but it asks for automatic monitoring and Feishu report delivery without clear user consent or scope controls.
Review before installing. Use it only with an intentionally configured browser profile, require confirmation before Feishu sends, and disable or tightly define Nasdaq and four-hour automatic triggers. Treat its investment conclusions as research notes, not trading advice.
The declared description presents a full Trump public-opinion analysis capability with multi-channel data acquisition and market-validation logic. The supplied code chunk does not implement any of those behaviors; it merely reads a specific local JSON file and prints basic summary information. This is a materially different and much narrower behavior than the declared purpose, so it should be flagged as a mismatch.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
[
{
"pubDate": "Wed, 01 Apr 2026 16:29:10 GMT",
"title": "US to leave Iran \u0027pretty quickly\u0027 and return if needed, Trump tells Reuters - Reuters",
"description": "US to leave Iran \u0027pretty quickly\u0027 and return if needed, Trump tells Reuters\u0026nbsp;\u0026nbsp;Reuters",
"link": "https://news.google.com/rss/articles/CBMiuwFBVV95cUxOQkNNSHlXV2dyc2dEWng3bm84T3dGdEhpdTVqNFFoQXNydDhsUzdYRkVYT1FmNUsxQlhWR1p1bHFRendVVGFrb0RJdVB2Zjhydno2MHFId05HYmtKb1Vfei1Ha0dLYW0wWU8xSlRIRHU3dWRQSXN2OVFNT0VFTlIyaHEwYlUyR1BFWFV1VXp2bGlLcE9NZGJXbExUbzE2eHVldVFGWWVSSl9vX0NneURyeFp0WjFER0VzOFZ3?oc=5",
"source": "GoogleNews_TrumpIranNuclear"
},
{
"pubDate": "Thu, 02 Apr 2026 21:29:45 GMT",
"title": "FactChe
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
[
{
"pubDate": "Wed, 01 Apr 2026 16:29:10 GMT",
"title": "US to leave Iran \u0027pretty quickly\u0027 and return if needed, Trump tells Reuters - Reuters",
"description": "US to leave Iran \u0027pretty quickly\u0027 and return if needed, Trump tells Reuters\u0026nbsp;\u0026nbsp;Reuters",
"link": "https://news.google.com/rss/articles/CBMiuwFBVV95cUxOQkNNSHlXV2dyc2dEWng3bm84T3dGdEhpdTVqNFFoQXNydDhsUzdYRkVYT1FmNUsxQlhWR1p1bHFRendVVGFrb0RJdVB2Zjhydno2MHFId05HYmtKb1Vfei1Ha0dLYW0wWU8xSlRIRHU3dWRQSXN2OVFNT0VFTlIyaHEwYlUyR1BFWFV1VXp2bGlLcE9NZGJXbExUbzE2eHVldVFGWWVSSl9vX0NneURyeFp0WjFER0VzOFZ3?oc=5",
"source": "GoogleNews_TrumpIranNuclear"
},
{
"pubDate": "Thu, 02 Apr 2026 21:29:45 GMT",
"title": "FactChe
The skill instructs browser-based collection from external services and the analyzer detected code-capable behavior such as file_read, but the manifest does not declare tool scope or permissions. Missing explicit scope weakens user/admin visibility into what the skill may access and increases the chance of over-privileged execution or accidental local data exposure when the skill runs in a broader-capability environment.
The manifest description is written entirely in Chinese and presents the skill as a Chinese-language experience, with no indication that users may choose another language or locale. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
The skill auto-activates on broad Nasdaq movement criteria and instructs immediate execution in the same session, which can trigger unsupervised browsing and analysis without a clear user action boundary. In this skill's context, that matters more because the workflow also includes external data collection and outbound Feishu reporting, so a vague trigger can cause unintended surveillance-like behavior or excessive automation.
The workflow says results may be pushed to Feishu automatically, but the skill does not clearly warn the user or obtain explicit consent before outbound notification. This creates a real data-leak risk because collected content, derived analysis, timestamps, and potentially user-trigger context could be transmitted to an external system without the user's awareness.
This JSON file includes natural-language output that is fixed to Chinese: "当前无强烈单一方向信号，建议观望等待明确催化剂". Because the file provides no user opt-in, language selection, or documented region-specific justification, it violates the policy against forcing a specific language or locale.
The file’s human-readable fields and narrative content are consistently written in Chinese, including status strings, notes, summaries, and key narratives. For a generally applicable intelligence/timeline data artifact, this constitutes a locale restriction with no visible opt-in, alternative language support, or documented justification in the file.
The manifest-like JSON includes a trigger value of "定时检查（每4小时）" without clarifying what system schedules it, what exact conditions cause execution, or any scope limitations or exclusions. For manifest/config files, vague activation descriptions can lead to unintended or misunderstood invocations because the trigger is described only broadly as a periodic check.
The manifest-style JSON includes a trigger value of "forced_timestamp_timeline", but it does not explain what conditions cause activation or what inputs are excluded. Without scope, constraints, or negative examples, the trigger semantics are ambiguous and could lead to unintended invocation behavior.
This JSON file consistently pairs English titles with Chinese description text in the "desc" fields, but provides no natural-language indication that the skill is intended only for Chinese-speaking users or that users can choose a language. Under the policy, forcing a specific language without user opt-in or justification is a locale-policy concern.
The natural-language interpretation is written in Chinese while the rest of the file content is predominantly English, and the file provides no indication that the user opted into Chinese output or can choose a locale. This creates a language/locale policy concern because the skill appears to force a specific language for part of its user-facing output.
The JSON content indicates a forced execution context for a Trump sentiment workflow, and the associated skill metadata uses broad trigger phrases such as '/trump' and '分析Trump舆情'. In a chat or agent environment, overly generic activation text can cause unintended invocation on incidental mentions of Trump, leading the agent to run a high-impact market/political analysis flow without clear user intent.
The file content is entirely structured around Chinese-language output and summaries, and the skill description is also Chinese-centric, with no indication of adapting to the user's preferred language. Forcing a non-user-selected language can degrade transparency and user comprehension, which is especially problematic in a skill influencing investment decisions based on geopolitical events.
This JSON file contains bilingual natural-language fields where snippets are consistently presented in Chinese while titles are in English, with no indication that Chinese output is optional or required for a specific region. That can violate language/locale policy because it imposes a language format on downstream users or systems without explicit opt-in or justification.
The JSON entries consistently pair English headlines with Chinese snippet text, indicating a fixed language behavior. Because this file provides no user choice, opt-in, or justification for the locale constraint, it appears to violate the language/locale policy requirement.
The top-level interpretation string is written in Chinese, and several metadata values elsewhere in the file also use Chinese terms such as "高" and Chinese keywords. In a general analysis artifact, this implies a fixed language/locale choice without visible user opt-in or a documented region-specific constraint, which matches the language/locale policy violation criteria.
This JSON includes a locale-specific formatted timestamp field using Chinese text and a CST label, which indicates the skill output is tailored to a specific language/locale. Under the policy, forcing a specific language or locale without user choice or clear region-specific justification is a natural-language policy concern.
Multiple user-facing text fields such as content, event descriptions, market impact, and notes are presented only in Chinese. Because the file provides no indication that the user opted into Chinese or that the skill is restricted to a Chinese-language context, this appears to enforce a specific language.
The input dataset is materially misaligned with the skill’s stated Trump-focused sentiment-analysis purpose: most items concern Iran, oil, airlines, and general macro commentary, with only limited direct Trump relevance. In an investment-decision support context, this can mislead downstream analysis, cause incorrect sentiment attribution to Trump, and produce distorted trading or research outputs from irrelevant news.
The report source includes 'Iran ceasefire search', which is outside the declared Trump-focused sentiment-analysis scope. This kind of scope expansion is dangerous because it can silently broaden data collection and downstream analysis beyond user expectations and approved capability boundaries, increasing the risk of unauthorized monitoring or misuse of the skill for unrelated geopolitical intelligence gathering.
The trigger_source value includes a forced execution indicator ('forced=true') but does not specify who can invoke it, under what conditions, or what safeguards exist. In an investment-oriented sentiment skill, ambiguous trigger semantics can enable unauthorized or unintended runs, increasing the risk of false alerts, noisy market signals, or abuse of downstream messaging/reporting channels.
The trigger condition combines a market move threshold with a vague 'forced scheduled trigger' clause, making activation semantics unclear and potentially causing the skill to run outside intended scenarios. In an investment-oriented geopolitical sentiment skill, ambiguous triggering can lead to unnecessary or mistimed dissemination of high-impact analysis, increasing the risk of operator confusion or poor downstream decisions.
This markdown file presents all headings and analysis content in Chinese and does not indicate that the user can select another language. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.
No suspicious patterns detected.