Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Security audit
ru-text
Security checks for vulnerabilities and agentic risk
Overview
This is a disclosed Russian text editing and typography skill with broad automatic styling behavior but no evidence of code execution, data access, persistence, or exfiltration.
Install this if you want Russian prose and UI text to be consistently edited with these typography and style norms. Be aware that it is configured for implicit invocation and silent typography changes on Russian output, so users who need exact Russian text reproduction, machine-readable strings, or nonstandard style should say so explicitly.
Vulnerability Patterns
- Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
- Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
- Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
- Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
- Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Rogue AgentSelf-Modification, Session Persistence
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (23)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The trigger phrases include very common Russian editing requests such as 'проверь текст', 'поправь', and 'отредактируй', which are likely to match ordinary user prompts outside the user's explicit intent to invoke this skill. This creates overbroad activation that can silently steer or rewrite responses, causing unintended behavior and reducing user control over how content is processed.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The rule 'Typography silently on any Russian output' is extremely broad and applies even when the user did not request editing, which can alter meaning, formatting, quoted text handling, or downstream machine-consumed output. Silent always-on modification is risky because it bypasses consent and can interfere with other tasks, tools, or higher-priority instructions that expect faithful reproduction.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
The skill is presented as specifically Russian-language text handling ('ru-text' and 'Russian text quality') with no indication that users can choose another language or opt into this locale restriction. This is a natural-language policy concern because it imposes a language/locale constraint without documenting user choice or justification.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The skill is presented as "ru-text" with a description focused exclusively on Russian text quality, which suggests a language-specific constraint. There is no accompanying natural-language indication that users can opt into this locale or that the restriction is explicitly justified as a region- or language-specific tool.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 80% confidence
- Finding
The text describes the rules as a catalogue of AI-generated Russian prose and repeatedly frames the rule set as normative for Russian writing. Because this is a markdown file containing natural-language guidance, this can amount to a language-policy constraint without any opt-in or explicit statement that the skill is only for users intentionally working on Russian text.
- Content
Session Persistence
Medium
- Category
- Rogue Agent
- Confidence
- 60% confidence
- Finding
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
- Content
md | A junction inside a passage being copy-edited as a book. | Literary register (AD-17.3). | **Severity:** Low. Secondary signal in the **С — Structure** dimension. A single junction is worth a remark and no more; the signal is density, and the fix is a rewrite the author would have made anyway. Cannot trigger non-compensatory caps alone. **Acknowledged:** raised by the director, 29.07.2026, on reading a junction in this project's own
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 86% confidence
- Finding
The file is framed as a prescriptive catalog for Russian text only, beginning with 'Russian Text Anti-Patterns' and immediately listing Russian-language corrections. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale constraint is clearly documented and justified.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The file is explicitly framed as a business writing guide in Russian and all guidance assumes Russian as the required output language. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is clearly documented and justified as region-specific.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
The document states prescriptive language behavior for Russian (for example, "Russian does NOT use Title Case") as a general rule, but does not clearly frame itself as applying only when the user is working in Russian. Under the policy, forcing a specific language/locale without opt-in or clear justified scope is a natural-language policy concern.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
Line L016 contains the example phrase "вызывай ru-text только когда я прошу" / "вызывай ru-text, только когда я прошу," which explicitly references invoking a Russian-language skill. In a policy/reference document, this can function as a locale-specific directive without offering opt-in or explaining that the skill is intentionally region- or language-specific.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
This file contains natural-language guidance that is effectively limited to one language and locale context. Under the policy rule, forcing a specific language without user opt-in or an explicit documented justification is a natural-language policy concern.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
Line L003 states the skill evaluates Russian text quality, and L025 further specifies input as Russian text of any length. This is a natural-language locale restriction, but the file does not present it as an opt-in choice or explain a region/compliance reason that would justify forcing a single language.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The title explicitly scopes the document to Russian typography rules, and multiple rules state that non-Russian punctuation conventions are unacceptable in Russian text. This is a natural-language locale constraint with no user opt-in or choice mechanism described in the file.
- Content
Ae4
Medium
- Category
- analysis-evasion
- Confidence
- 80% confidence
- Finding
Suspicious Unicode normalization or mixed-script content
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 86% confidence
- Finding
The title and description explicitly frame the guidance as "UX-writing in Russian" and present it as a general reference, which can amount to a language-policy constraint. Under the policy rule, forcing a specific language without opt-in or clear region-specific justification is a natural-language policy violation.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 83% confidence
- Finding
The text states that the referenced works have transformed the culture of Russian text on the internet, and the cited sources throughout the file are exclusively Russian-language style and typography references. Because SQP-3 applies to language/locale policy violations, this can be read as enforcing a Russian-language locale without any explicit user opt-in or alternative language choice.
- Content
Static analysis
No suspicious patterns detected.