Security audit

Summon Translator Skill

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent cloud translation integration, but its sample upload command can execute unintended local shell commands if given a crafted file path.

Install only if you are comfortable sending chosen documents to Summon Translator’s external service and using a paid API key. Treat the current shell examples cautiously: do not run them with untrusted file paths, and prefer a fixed script or API client that passes file paths as arguments instead of substituting them into shell text.

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

Error

Location: SKILL.md:73
Finding: Shell Command Injection Through Unquoted File Path
Content: View full analysis
/tmp/summon-command-output ``` 2. The agent substitutes that value directly into the documented command. 3. The final name argument becomes equivalent to: ```bash -F "name=$(basename document; id > /tmp/summon-command-output)" ``` 4. When the command is interpreted by a shell, `basename document` runs first, followed by the injected `id` command. 5. The injected command executes with the same operating-system identity and permissions ...[truncated 1038 chars]
Remediation: View remediation
&2 exit 1 fi curl -s -X POST 'https://summontranslator.com/api/v1/jobs' \ -H "Authorization: Bearer $SUMMON_API_KEY" \ -F "file=@${file_path}" \ -F "provider=${provider}" \ -F "model=${model}" \ -F "targetLanguages=${target_languages}" \ -F "sourceLanguage=${source_language}" \ -F "name=${output_name}" ``` The caller should invoke this script using a non-shell argument array so that the supplied path remains one argument regardless of embedded shell metacharacters. Apply the same structured argument handling to all other placeholders used by the documented commands. ]]>

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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 (1)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill instructs users to upload local files to summontranslator.com via curl, but it does not present a clear upfront warning that file contents will be transmitted to a third-party external service. Because the skill is explicitly aimed at compliance, supplier, and safety documents—materials that often contain sensitive operational, legal, or personal data—users may disclose confidential content without informed consent.
Content: No source excerpt is available for this finding.