Security audit
Security checks for vulnerabilities and agentic risk
The skill is a static Michael Burry investing-perspective reference with no executable behavior, but users should treat its market examples cautiously.
Install this only if you want a Burry-style investing lens. Treat all stock, market, short-selling, and portfolio examples as informational and historical/contextual, verify current facts independently, and do not rely on this skill for personalized financial advice or trading decisions.
The trigger list contains broad, generic investing terms like 'value investing,' 'margin of safety,' and 'housing market' that are not uniquely tied to this skill. This can cause the agent to invoke the Burry persona in unrelated finance discussions, leading to unintended behavioral steering and lower reliability of responses.
The trigger list contains broad phrases such as "value investing," "contrarian investing," "margin of safety," and "short selling" that are common financial discussion terms rather than uniquely identifying this skill. This can cause unintended activation in unrelated investing conversations, leading the agent to inject specialized guidance without clear user intent and increasing the chance of misrouting or inappropriate financial advice.
The single-word trigger "Cassandra" is highly ambiguous and overlaps with common personal names, mythology, software products, and other unrelated topics. This makes accidental invocation especially likely, which can cause the skill to activate in contexts having nothing to do with Michael Burry or investing.
This section presents concrete market theses, named securities, bearish positions, and a paid subscription pitch in a skill intended to shape user decision-making, but it does not include any disclaimer that the material is informational rather than personalized financial advice. In an agent skill context, this increases the chance that users treat speculative commentary as actionable guidance, potentially leading to financial harm or inappropriate reliance on the model for regulated advice-like outputs.
This file presents detailed, current-seeming trading decisions, concentrated portfolio allocations, hedging positions, and celebrated outcomes without any disclaimer that the material is educational or historical rather than investment advice. In an agent skill framed as a reusable cognitive operating system, users could reasonably treat these examples as actionable recommendations, increasing the risk of harmful financial decisions.
The phrase '二手 analysis' introduces mixed-language content without any indication that multilingual behavior is expected. In practice this can confuse parsing, degrade user comprehension, or produce inconsistent outputs if the agent mirrors or adopts unexpected language in its responses.
The manifest sets the language to "en," which indicates an English-only configuration, but there is no accompanying justification that this skill is region-specific or any indication of user opt-in for language preference. Under the policy, forcing a specific language without choice or documented justification can be a natural-language policy issue.
Lines L073 and L078 describe events extending into October/November 2025 in a definitive tone, and the document is marked last updated in 2026. This is a natural-language content issue because the file presents asserted factual claims without any indication of uncertainty or user choice, which can violate organizational expectations for accurate, policy-compliant reference material.
No suspicious patterns detected.
