kodagpt

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the membrane CLI to manage authentication, connections, and data operations with KodaGPT. These commands are part of the intended integration workflow for the vendor's platform.- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli package from the npm registry to enable interaction with the service. This is a standard installation of a vendor-provided tool.- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes content from KodaGPT knowledge bases and chatbot messages.
  • Ingestion points: Data is ingested through actions such as list-contents, semantic-search, and send-message-to-chatbot in SKILL.md.
  • Boundary markers: No explicit delimiters or boundary markers are used in the provided command examples to isolate untrusted data from the agent's instructions.
  • Capability inventory: The skill possesses capabilities to create, update, and delete data using membrane action run and can perform arbitrary network requests via membrane request.
  • Sanitization: The instructions do not specify any sanitization or validation logic for the data retrieved from external API calls.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 10:52 PM