didit-sdk-integration

Pass

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill includes a detailed prompt template under the 'Three-product integration prompt' section designed to take user-supplied application context (e.g., Stack, Surface, Products, Use case) and use it to generate integration code.\n
  • Ingestion points: The prompt template in SKILL.md is the primary ingestion point for untrusted user data via placeholders such as [framework and language] and [describe].\n
  • Boundary markers: The template uses square brackets [...] to denote where user input should be placed.\n
  • Capability inventory: The skill facilitates network operations (API calls to vendor domains) and local command execution (SDK installation, running helper scripts).\n
  • Sanitization: No explicit sanitization or filtering of the interpolated user input is described in the prompt template itself, potentially allowing the user to inject instructions into the agent's task.\n- [EXTERNAL_DOWNLOADS]: The skill references multiple external software packages and resources required for integration.\n
  • Evidence: Mentions NPM packages (@didit-protocol/sdk-web, @didit-protocol/sdk-react-native), Maven artifacts (me.didit:didit-sdk), and Swift packages hosted on GitHub (github.com/didit-protocol/sdk-ios.git).\n
  • Details: All identified resources originate from the vendor's own infrastructure or official package registries.\n- [COMMAND_EXECUTION]: The skill provides examples of shell commands for installing dependencies and running a utility script.\n
  • Evidence: npm install, flutter pub add, and python scripts/mint_sdk_tokens.py commands are provided in SKILL.md and the script's docstrings.\n
  • Details: These commands are part of the intended developer workflow for SDK integration and utility execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 14, 2026, 10:28 AM