Security audit
Security checks for vulnerabilities and agentic risk
This is a content-marketing guidance skill with no executable code, data access, persistence, or hidden high-impact behavior found.
Install only if you want an opinionated content-marketing assistant, including Web3/crypto examples. Be aware it may activate on broad content or crypto mentions and may answer partly in Spanish unless you specify your preferred language.
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<rect width="680" height="420" fill="white" rx="8"/>
<!-- Top accent bar -->
<rect x="0" y="0" width="680" height="4" fill="#1D9E75" rx="0"/>
<!-- Grid lines -->
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<!-- Divider -->
<line x1="40" y1="152" x2="640" y2="152" stroke="#e5e5e5" stroke-width="0.5"/>
<!-- Badges -->
<rect x="40" y="164" width="102" height="24" rx="4" fill="#E1F5EE" stroke="#0F6E56" stroke-width="0.5"/>
<text x="91" y="181" text-anchor="middle" font-family="ui-sans-serif, system-ui, sans-serif" font-size="11px" font-weight="500" fill="#085041">Web3 / crypto</text>
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
<text x="444" y="262" font-family="ui-sans-serif, system-ui, sans-serif" font-size="12px" fill="#185FA5">Hooks, posts, captions</text>
<text x="444" y="276" font-family="ui-sans-serif, system-ui, sans-serif" font-size="12px" fill="#185FA5">per platform</text>
<!-- Row 2 cards -->
<rect x="40" y="300" width="182" height="68" rx="6" fill="#FAEEDA" stroke="#854F0B" stroke-width="0.5"/>
<text x="56" y="324" font-family="ui-sans-serif, system-ui, sans-serif" font-size="14px" font-weight="500" fill="#412402">Repurposing Engine</text>
<text x="56" y="342" font-family="ui-sans-serif, system-ui, sans-serif" font-size="12px" fill="#854F0B">1 piece → 9 formats,</text>
The README states the skill activates automatically on broad keyword mentions like crypto, Web3, DeFi, NFTs, or blockchain, which can cause the skill to trigger outside the user's intended task scope. This is not inherently malicious, but it increases the chance of inappropriate routing, irrelevant prompt injection surface, or accidental use of a specialized module when the user only mentions the topic in passing.
The trigger phrases are broad enough to match generic requests like 'help me with my content' or 'write a post for me,' which can cause the skill to activate outside the user's actual intent. This is dangerous because unintended invocation can override more appropriate skills, degrade routing accuracy, and lead to irrelevant or misleading outputs in multi-skill environments.
The skill mandates a Spanish prompt for tone selection without checking the user's preferred language. This can create unexpected language switching, confuse users, and produce unusable output for English-only interactions, which is a quality and safety issue in agent behavior.
These instructions repeatedly require Spanish-language output elements and examples without offering locale choice or fallback behavior. Repeated forced language behavior increases the chance of non-compliant, confusing, or inaccessible responses, especially when the surrounding skill is otherwise written for general use.
The instruction to load this file for 'any user operating in crypto, Web3, DeFi, NFTs, blockchain, or fintech' is very broad and lacks boundaries or exclusion conditions. Terms like 'fintech' and 'operating in' are expansive enough to overlap with many general business or technology contexts, increasing the chance of unintended invocation.
The audience table specifies 'Spanish-speaking LATAM' and directs a particular tone and framing for that group, and a later section reinforces a 'Spanish-language crypto education' regional voice. Because the file presents this as prescribed language behavior without mentioning user preference or opt-in, it creates a potential language/locale policy concern.
No suspicious patterns detected.