cxas-composite-voice-agent-optimizer

Warn

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The scripts/audit_agent.py script uses importlib.util to dynamically load the linter.py module from a computed filesystem path. This is a technique for resolving local dependencies at runtime but is classified as dynamic code loading.
  • [COMMAND_EXECUTION]: The skill requires the execution of shell commands and the cxas CLI tool to perform audits, remediate local configuration files, and sync with cloud resources.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes natural language instruction files, which presents an attack surface where untrusted content in those files could influence the audit or remediation logic.
  • Ingestion points: Reads global_instruction.txt and agents/*/instruction.txt files from the local workspace.
  • Boundary markers: No explicit markers are used to isolate audited instruction content from the execution logic.
  • Capability inventory: Performs file modifications and executes shell commands via the CLI.
  • Sanitization: Processing relies on regular expressions and structural JSON validation, lacking isolation for raw natural language content.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 16, 2026, 02:19 PM
Security Audit — agent-trust-hub — cxas-composite-voice-agent-optimizer