translator

• Outcome: Pass
• Latest audit: Sep 11, 2026 · 3:17 AM UTC
• Version: 1.0.0

Overview

This is a coherent translation helper with broad but expected translation triggers and no code, dependencies, persistence, credential access, or hidden execution behavior.

Review outputs carefully for legal, medical, financial, or publication-critical translations, especially because this is not a certified human translator. The broad triggers may activate for ordinary translation requests, but the inspected package does not show hidden execution, credential handling, persistence, or data mutation.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (9)

Anti-Refusal Statement

High

• Category: Anti-Refusal

Content

- **Fix**: Identify all placeholders, tags, and markup before translating; preserve them exactly; only translate the natural-language text surrounding them

### 16. Ignoring String Length Constraints
- **Problem**: Producing translations that are too long for UI elements (buttons, tooltips, menu items) without warning
- **Example**: EN button "Submit" (6 chars) --> DE "Einreichen" (10 chars) or FR "Soumettre" (9 chars) — may overflow the button
- **Fix**: Check expansion/contraction rates from knowledge/domain.md; flag translations that exceed typical length budgets; offer abbreviated alternatives

• Confidence: 80% confidence
• Finding: Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Credential Access

High

• Category: Privilege Escalation

Content

"id": "bench-med-04",
      "difficulty": "medium",
      "description": "Translate a technical document with placeholders and code from English to Korean",
      "input": "Translate this API documentation excerpt into Korean. Preserve all code examples, variable names, and placeholders exactly as they appear. Translate only the natural-language descriptions.\n\n\"## Authentication\n\nAll API requests require a valid access token. Include the token in the `Authorization` header:\n\n```\nAuthorization: Bearer {access_token}\n```\n\n### Getting a Token\n\nSend a POST request to `/api/v2/auth/token` with your credentials:\n\n```json\n{\n  \"client_id\": \"your_client_id\",\n  \"client_secret\": \"your_client_secret\",\n  \"grant_type\": \"client_credentials\"\n}\n```\n\n**Response:** Returns a JSON object containing `access_token` (string), `expires_in` (integer, seconds), and `token_type` (always `\"Bearer\"`).\n\n**Note:** Tokens expire after 3600 seconds (1 hour). Implement token refresh logic to avoid service interruptions.\"",
      "rubric": [
        {
          "criterion": "Placeholder Preservation",

• Confidence: 70% confidence
• Finding: Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Natural-Language Policy Violations

Medium

• Confidence: 81% confidence
• Finding: The document presents language-pair guidance beginning with English-to-Chinese register signals using prescriptive language, which can be read as a default language/locale preference rather than an optional example set. Under the policy, forcing a specific language or locale without explicit user choice can be a natural-language policy violation.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The trigger set is broad and includes common natural-language phrases such as 'translate', 'say this in', and 'how do you say', which are likely to appear in ordinary conversation. This increases the chance of unintended or overly eager skill activation, potentially causing the agent to switch behavior unexpectedly and process user content under the skill's instructions when the user did not intend that mode.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: This JSON manifest contains many generic task descriptions and inputs such as 'Translate this...' across multiple languages, but it does not define any activation boundaries, explicit trigger lists, or negative examples. In a manifest-like file, such broad phrases overlap with common user requests and could cause unintended invocation if used as skill triggers or routing descriptions.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: Multiple tasks explicitly require outputs in Spanish, French, German, Simplified Chinese, Japanese, Arabic, Korean, English, and Brazilian Portuguese. Because the file hard-codes language/locale requirements as instructions rather than presenting them as user-selectable options or documenting a justified regional scope, it creates a natural-language locale policy concern.

Natural-Language Policy Violations

Medium

• Confidence: 93% confidence
• Finding: The task explicitly instructs translation into Simplified Chinese, which imposes a specific language/locale requirement. For this file type, natural-language policy issues apply, and there is no indication that the language choice is user-selected or that the constraint is justified as a region-specific requirement.

Overly Broad Trigger

Low

• Category: Trigger Abuse
• Confidence: 70% confidence
• Finding: Overly Broad Trigger: '翻译' is too short and may match unintended inputs

Scope Creep

Low

• Category: Excessive Agency

Content

"id": "bench-med-01",
      "difficulty": "medium",
      "description": "Translate a legal clause from English to Chinese with formal equivalence",
      "input": "Translate this legal clause into Simplified Chinese. This is from a commercial contract and must preserve precise legal meaning. Use formal equivalence approach.\n\n\"Limitation of Liability. In no event shall either party be liable to the other for any indirect, incidental, special, consequential, or punitive damages, including but not limited to loss of profits, data, or business opportunities, arising out of or related to this Agreement, regardless of the theory of liability (contract, tort, strict liability, or otherwise), even if such party has been advised of the possibility of such damages. Each party's total cumulative liability under this Agreement shall not exceed the fees paid by Client to Provider during the twelve (12) month period immediately preceding the event giving rise to such liability.\"",
      "rubric": [
        {
          "criterion": "Legal Accuracy",

• Confidence: 70% confidence
• Finding: Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Static analysis

No suspicious patterns detected.