ClawHub
Back to skill

Security audit

Magyar Szótár

Security checks for vulnerabilities and agentic risk

Overview

This is a small Hungarian dictionary/reference skill with disclosed public dictionary lookup fallback and no executable code, persistence, or privilege-seeking behavior.

Install only if you want Hungarian dictionary assistance. Avoid using external fallback on confidential documents unless you are comfortable sending lookup terms or batches to public dictionary services.

A.I.G

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:47
Finding

Uncontrolled disclosure of user-provided text to third-party dictionary services

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 47-75
Vulnerability Type: Unrestricted third-party data transmission
Risk Level: Medium

Vulnerable Code Snippet:

markdown
### 2. Web API fallback

Ha helyi szótár nem elérhető, használj nyilvános API-t:
- dict.sztaki.hu (magyar-angol, de van értelmező is)
- Wiktionary API

## Gyors kezdés

```python
# Példa szójelentés lekérdezésre
def lookup_word(word):
    # 1. Helyi szótár keresés
    if word in local_dict:
        return local_dict[word]
    # 2. API fallback
    return fetch_from_api(word)

Referenciák

  • references/kozmondasok.md – 100+ magyar közmondás és szólás jelentéssel, példákkal
  • references/magyar-szotar.json – fő szótár adatbázis
  • references/arcanum-ertsz.md – Arcanum ÉrtSz. forrásdokumentáció
  • references/morfológia.md – magyar nyelvtani szabályok, ragozások

Jegyzetek

  • Mindig ellenőrizd a szótár forrás frissességét
  • Nagy szövegeknél batch lookup-ot használj (token hatékony)
  • Ha szó nem található, jelezd és ajánlj alternatívát
text

### Technical Analysis

The Skill instructs the agent to use public web APIs whenever the local dictionary does not contain a requested word. The bundled local dictionary contains only four entries, so external fallback is likely during ordinary use.

The instructions do not require the agent to:

- Obtain explicit user consent before making a network request.
- Determine whether the submitted text is confidential or contains personal information.
- Restrict transmission to a single normalized dictionary word.
- Redact names, identifiers, credentials, or other sensitive material.
- Use a fixed allowlist of exact HTTPS API endpoints.
- Explain third-party logging, retention, or privacy implications.
- Remain in local-only mode when handling sensitive documents.

The recommendation to perform batch lookups for large texts increases th
...[truncated 1811 chars]
Remediation
View remediation

Remediation Suggestions

  1. Default to local-only processing and treat network fallback as an optional capability.
  2. Obtain explicit user consent before sending any lookup data to a third party.
  3. Tell the user which service will receive the data and what exact terms will be transmitted.
  4. Send only individually normalized unknown words; never transmit complete sentences, paragraphs, documents, or unreviewed batches.
  5. Disable external lookup automatically when input may contain credentials, personal data, confidential business information, medical information, or legal material.
  6. Replace the generic fetch_from_api behavior with a fixed allowlist of exact HTTPS endpoints and enforce certificate validation, request timeouts, response-size limits, and safe response parsing.
  7. Provide a local-only failure response when consent is absent, such as reporting that the term is not available in the bundled dictionary.
  8. Expand the bundled dictionary or use an approved offline linguistic database to reduce unnecessary external requests.
  9. Document applicable third-party privacy, retention, and logging policies.
  10. Remove or strictly constrain batch lookup behavior so that batches cannot reconstruct meaningful portions of the source text.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (4)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest description and usage section are entirely scoped to Hungarian requests and instruct the agent to handle only Hungarian word meanings, synonyms, and analysis. This constitutes a language/locale constraint in natural language without any explicit user opt-in or statement that the skill is region- or language-specific by design.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This file presents all instructions, metadata, and examples exclusively in Hungarian, which can amount to a language/locale policy violation when no user opt-in or alternative language is offered. The content does not state that the skill is intentionally limited to Hungarian users or a region-specific compliance context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file presents all headings, descriptions, and usage guidance exclusively in Hungarian. Under the policy for natural-language violations, forcing a specific language without user opt-in or a documented justification can exclude users and violates locale-choice expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This JSON file is entirely described and populated in Hungarian, including the metadata description and example content, with no indication that the user can choose another language or locale. Under the natural-language policy rule, forcing a specific language without opt-in can be a policy concern unless the locale restriction is explicitly documented and justified.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.