🛠️
Audit-Verlauf
ai-runtime-toolkit - 5 Audits
Audit-Version 5
Neueste Mittleres RisikoJan 17, 2026, 04:30 AM
The toolkit provides legitimate development utilities with tool discovery and management capabilities. Several input validation issues were identified in bash scripts and Python CLI that require user awareness. No malicious behavior or credential exfiltration detected. Risk level set to medium due to command injection potential in tool execution paths.
59
Gescannte Dateien
7,409
Analysierte Zeilen
9
befunde
claude
Auditiert von
Probleme mit hohem Risiko (3)
Command Injection in CLI Runner
discover/cli.py:177-178 passes unsanitized user arguments to subprocess.run
Command Injection in Log Analyzer
bash/analysis/analyze-logs.sh:84,88 uses unsanitized user input in grep patterns
Unsourced Environment Loading
bash/system/check-service.sh:34-35 sources .env without validation
Probleme mit mittlerem Risiko (1)
Path Traversal Potential
Python tools traverse directories without strict path validation
Probleme mit niedrigem Risiko (1)
File Reading Without Strict Validation
Dependency analyzer reads package files without path validation
Risikofaktoren
⚙️ Externe Befehle (2)
📁 Dateisystemzugriff (2)
🌐 Netzwerkzugriff (2)
⚡ Enthält Skripte (2)
Erkannte Muster
Direct Shell Command Execution with User Input
Audit-Version 4
Mittleres RisikoJan 17, 2026, 04:30 AM
The toolkit provides legitimate development utilities with tool discovery and management capabilities. Several input validation issues were identified in bash scripts and Python CLI that require user awareness. No malicious behavior or credential exfiltration detected. Risk level set to medium due to command injection potential in tool execution paths.
59
Gescannte Dateien
7,409
Analysierte Zeilen
9
befunde
claude
Auditiert von
Probleme mit hohem Risiko (3)
Command Injection in CLI Runner
discover/cli.py:177-178 passes unsanitized user arguments to subprocess.run
Command Injection in Log Analyzer
bash/analysis/analyze-logs.sh:84,88 uses unsanitized user input in grep patterns
Unsourced Environment Loading
bash/system/check-service.sh:34-35 sources .env without validation
Probleme mit mittlerem Risiko (1)
Path Traversal Potential
Python tools traverse directories without strict path validation
Probleme mit niedrigem Risiko (1)
File Reading Without Strict Validation
Dependency analyzer reads package files without path validation
Risikofaktoren
⚙️ Externe Befehle (2)
📁 Dateisystemzugriff (2)
🌐 Netzwerkzugriff (2)
⚡ Enthält Skripte (2)
Erkannte Muster
Direct Shell Command Execution with User Input
Audit-Version 3
Mittleres RisikoJan 10, 2026, 02:35 PM
The toolkit system has several input validation vulnerabilities that could allow command injection and path traversal attacks. While not inherently malicious, these issues need addressing before production use.
49
Gescannte Dateien
6,499
Analysierte Zeilen
10
befunde
claude
Auditiert von
Probleme mit hohem Risiko (2)
Command Injection in Log Analyzer
The analyze-logs.sh script directly embeds user input in grep commands without sanitization. Lines 84 and 88 use: grep "$PATTERN" "$LOG_FILE". An attacker could inject malicious commands by providing patterns like '; rm -rf / #' in the PATTERN variable, leading to arbitrary command execution.
Unvalidated Subprocess Execution
The CLI runner executes tools via subprocess.run() without validating the tool path or arguments. Line 177-178: cmd = [str(tool_path)] + args.args. If an attacker can manipulate the tool discovery system or file system, they could execute arbitrary commands.
Probleme mit mittlerem Risiko (3)
Path Traversal in File Processing Tools
Multiple Python tools perform recursive directory traversal without validating paths. The code-stats.py tool (lines 95-108) follows symbolic links and doesn't validate file paths, potentially allowing access to sensitive files outside the intended directory.
SSRF Risk in API Tester
The test-api.js tool accepts a baseURL parameter without validation. Lines 150-170 use node-fetch with user-provided URLs, which could be exploited for Server-Side Request Forgery attacks to access internal services.
Unvalidated Environment Loading
The check-service.sh script sources .env files without validation (line 35: source .env). Malicious environment variables could be loaded, potentially affecting system behavior or exposing sensitive data.
Probleme mit niedrigem Risiko (1)
File Reading Without Validation
Dependency analyzer reads package.json and requirements.txt files without validating paths or checking file sizes. Could be exploited to read arbitrary files or cause DoS with large files.
Risikofaktoren
⚡ Enthält Skripte (3)
📁 Dateisystemzugriff (3)
⚙️ Externe Befehle (2)
🌐 Netzwerkzugriff (2)
Erkannte Muster
Direct Shell Command Execution
Audit-Version 2
Mittleres RisikoJan 10, 2026, 02:35 PM
The toolkit system has several input validation vulnerabilities that could allow command injection and path traversal attacks. While not inherently malicious, these issues need addressing before production use.
49
Gescannte Dateien
6,499
Analysierte Zeilen
10
befunde
claude
Auditiert von
Probleme mit hohem Risiko (2)
Command Injection in Log Analyzer
The analyze-logs.sh script directly embeds user input in grep commands without sanitization. Lines 84 and 88 use: grep "$PATTERN" "$LOG_FILE". An attacker could inject malicious commands by providing patterns like '; rm -rf / #' in the PATTERN variable, leading to arbitrary command execution.
Unvalidated Subprocess Execution
The CLI runner executes tools via subprocess.run() without validating the tool path or arguments. Line 177-178: cmd = [str(tool_path)] + args.args. If an attacker can manipulate the tool discovery system or file system, they could execute arbitrary commands.
Probleme mit mittlerem Risiko (3)
Path Traversal in File Processing Tools
Multiple Python tools perform recursive directory traversal without validating paths. The code-stats.py tool (lines 95-108) follows symbolic links and doesn't validate file paths, potentially allowing access to sensitive files outside the intended directory.
SSRF Risk in API Tester
The test-api.js tool accepts a baseURL parameter without validation. Lines 150-170 use node-fetch with user-provided URLs, which could be exploited for Server-Side Request Forgery attacks to access internal services.
Unvalidated Environment Loading
The check-service.sh script sources .env files without validation (line 35: source .env). Malicious environment variables could be loaded, potentially affecting system behavior or exposing sensitive data.
Probleme mit niedrigem Risiko (1)
File Reading Without Validation
Dependency analyzer reads package.json and requirements.txt files without validating paths or checking file sizes. Could be exploited to read arbitrary files or cause DoS with large files.
Risikofaktoren
⚡ Enthält Skripte (3)
📁 Dateisystemzugriff (3)
⚙️ Externe Befehle (2)
🌐 Netzwerkzugriff (2)
Erkannte Muster
Direct Shell Command Execution
Audit-Version 1
Mittleres RisikoJan 10, 2026, 02:35 PM
The toolkit system has several input validation vulnerabilities that could allow command injection and path traversal attacks. While not inherently malicious, these issues need addressing before production use.
49
Gescannte Dateien
6,499
Analysierte Zeilen
10
befunde
claude
Auditiert von
Probleme mit hohem Risiko (2)
Command Injection in Log Analyzer
The analyze-logs.sh script directly embeds user input in grep commands without sanitization. Lines 84 and 88 use: grep "$PATTERN" "$LOG_FILE". An attacker could inject malicious commands by providing patterns like '; rm -rf / #' in the PATTERN variable, leading to arbitrary command execution.
Unvalidated Subprocess Execution
The CLI runner executes tools via subprocess.run() without validating the tool path or arguments. Line 177-178: cmd = [str(tool_path)] + args.args. If an attacker can manipulate the tool discovery system or file system, they could execute arbitrary commands.
Probleme mit mittlerem Risiko (3)
Path Traversal in File Processing Tools
Multiple Python tools perform recursive directory traversal without validating paths. The code-stats.py tool (lines 95-108) follows symbolic links and doesn't validate file paths, potentially allowing access to sensitive files outside the intended directory.
SSRF Risk in API Tester
The test-api.js tool accepts a baseURL parameter without validation. Lines 150-170 use node-fetch with user-provided URLs, which could be exploited for Server-Side Request Forgery attacks to access internal services.
Unvalidated Environment Loading
The check-service.sh script sources .env files without validation (line 35: source .env). Malicious environment variables could be loaded, potentially affecting system behavior or exposing sensitive data.
Probleme mit niedrigem Risiko (1)
File Reading Without Validation
Dependency analyzer reads package.json and requirements.txt files without validating paths or checking file sizes. Could be exploited to read arbitrary files or cause DoS with large files.
Risikofaktoren
⚡ Enthält Skripte (3)
📁 Dateisystemzugriff (3)
⚙️ Externe Befehle (2)
🌐 Netzwerkzugriff (2)
Erkannte Muster
Direct Shell Command Execution