capacitor-plugin-spm-support

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill incorporates external code and repositories into the project's dependency graph.
  • It explicitly references https://github.com/ionic-team/capacitor-swift-pm.git as a dependency in the generated Package.swift file.
  • It instructs the agent to search the web for third-party Swift Package Manager (SPM) equivalents for existing CocoaPods dependencies and add their repository URLs to the manifest file.
  • [COMMAND_EXECUTION]: The skill executes shell commands to install and resolve dependencies.
  • It prompts the agent to run npm install within the plugin root directory to verify changes.
  • It adds and executes a new script entry in package.json: "ios:spm:install": "cd ios && swift package resolve && cd ..".
  • [DYNAMIC_EXECUTION]: The skill dynamically generates configuration files and scripts that are subsequently executed by system tools.
  • It constructs a Package.swift manifest from scratch using information gathered from the plugin environment and external searches.
  • It injects new command-line scripts into the existing package.json file.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection through the processing of untrusted project files and external search data.
  • Ingestion points: Reads local files including package.json, .podspec, and Swift source files (ios/Plugin/*.swift). It also ingests URLs and version requirements retrieved from web search results.
  • Boundary markers: The instructions lack explicit boundary markers or delimiters to isolate the content of processed files from the agent's internal logic.
  • Capability inventory: The skill has the capability to write to the file system (Package.swift, package.json, .gitignore, .xcodeproj), execute shell commands (npm install), and trigger automated package resolution (swift package resolve).
  • Sanitization: There is no evidence of sanitization or validation logic to ensure that extracted values (such as version strings, package names, or external repository URLs) do not contain malicious payloads designed to manipulate the generated code or command execution flow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 05:21 PM