Security audit
Security checks for vulnerabilities and agentic risk
This is a text-only Russian editing skill whose broad automatic activation may affect Russian wording and typography, but I found no hidden code, data access, persistence, or destructive behavior.
Install this if you want Russian text produced or edited by your agent to follow these typography and style rules automatically. Avoid it if you need Russian output preserved exactly as typed unless you explicitly instruct the agent to preserve formatting or quoted material.
The trigger scope is excessively broad: it includes common phrases like «проверь текст» and an especially risky catch-all condition, «Also any Russian output.» This can cause the skill to activate on a large share of unrelated Russian-language requests, unexpectedly overriding normal agent behavior and creating prompt-routing hijack conditions.
Suspicious Unicode normalization or mixed-script content
The skill states ‘Apply these rules to ALL Russian text output without exception,’ which forces a language/transformation policy even when the user did not request editing or typography normalization. This can silently alter meaning, formatting, quoted material expectations, or downstream output requirements, making the skill intrusive when combined with its broad activation rules.
The manifest identifies the skill as "ru-text" and describes it exclusively as "Russian text quality," which indicates a language-specific constraint. There is no surrounding text indicating that Russian output is optional, user-selected, or otherwise justified as a region-specific tool.
The metadata presents the skill as Russian-text oriented and the broader skill context indicates activation for 'any Russian output,' but there is no indication of user choice or opt-in. This can override user expectations, cause unwanted language/style transformations, and combine with implicit invocation to make the skill trigger too broadly.
The skill enables implicit invocation broadly without any trigger constraints or activation examples, which can cause it to activate on general Russian output or loosely related editing requests. That increases the chance of unintended routing, silent behavior changes, or prompt-surface expansion, especially because the skill description says it applies to 'any Russian output.'
Suspicious Unicode normalization or mixed-script content
The document states that it is a catalogue of tells of AI-generated Russian prose and then frames the rule set around that language specifically. Under the policy rule, forcing a specific language/locale is acceptable only with clear documentation and justification; this file does not explicitly say the scope is intentionally limited to Russian for a region- or language-specific tool.
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.
| 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
Suspicious Unicode normalization or mixed-script content
The document is explicitly framed as "Russian Text Anti-Patterns" and all examples and corrections are presented as Russian-language norms. Under the policy, forcing a specific language without user opt-in or a clearly documented regional/compliance justification is a natural-language policy violation.
The title and opening describe the skill as a business writing guide specifically in Russian, but the document does not indicate that this language restriction is optional, user-selected, or limited to a clearly justified regional/compliance context. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern.
Suspicious Unicode normalization or mixed-script content
This file states prescriptive language rules for Russian usage ('Russian does NOT use Title Case') as mandatory guidance, and the broader document consistently enforces Russian grammar conventions. Under the policy, forcing a specific language or locale without user opt-in or clear justification is a natural-language policy concern.
The file consists of editorial rules and examples centered entirely on Russian punctuation and grammar, including explicit Russian conjunctions and examples. This imposes a specific language/locale context, but the document does not state that the skill is Russian-specific or offer any user choice, which can violate the policy against forcing a language without opt-in.
Suspicious Unicode normalization or mixed-script content
The description states that this is a reference specifically for Russian text, which imposes a language constraint in the skill content. Under the policy, language-specific behavior should either offer user choice or clearly justify the locale restriction; this file does not present such opt-in or justification.
Suspicious Unicode normalization or mixed-script content
The rubric explicitly defines its input as 'Russian text of any length,' which constrains language use. The document does not offer a user language choice or explain a justified region- or compliance-specific reason for enforcing Russian, so this is a natural-language locale policy concern.
Suspicious Unicode normalization or mixed-script content
The document title explicitly defines the skill content as rules for Russian typography, which imposes a specific language/locale context. Under the stated policy, locale-specific behavior should either offer user opt-in or clearly document a justified region-specific constraint; this file does not present that choice or justification in user-facing policy terms.
Suspicious Unicode normalization or mixed-script content
The title and description state that this guidance is specifically for UX writing in Russian, framing the skill around a single language/locale. 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 as region-specific.
The text describes the project as focused on "Russian text on the internet," and the referenced materials throughout the file are exclusively Russian-language sources. For a policy check on language/locale constraints, this can be read as enforcing a specific locale without an explicit opt-in or justification in the file.
No suspicious patterns detected.