check-saaskit-prod

Warn

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The instructions in Step 2 require the agent to execute echo $SCALEKIT_CLIENT_SECRET. This command causes a production secret to be printed to the standard output, which exposes sensitive credentials in the agent's interaction history and session logs.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands including echo, curl, and rg (ripgrep). The curl command is used to perform an OAuth client credentials grant flow, transmitting production secrets to the environment URL.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill uses the rg command to scan the local codebase (SKILL.md, Step 2) and return matching lines to the agent context.
  • Boundary markers: No delimiters or safety instructions are provided to prevent the agent from being influenced by potentially malicious strings found within the local files during the grep operation.
  • Capability inventory: The agent has capabilities to execute shell commands (curl, echo, rg) and perform network operations.
  • Sanitization: There is no filtering or sanitization of the data retrieved from the filesystem before it is processed by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 28, 2026, 08:07 AM
Security Audit — agent-trust-hub — check-saaskit-prod