ci-security-scanning-with-strix

Fail

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructions specify an installation method that downloads a shell script from the vendor's domain (https://strix.ai/install) and pipes it directly into bash. This is a vendor-controlled resource.
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations to download tools and interact with managed services on vendor-controlled domains, specifically strix.ai and app.strix.ai.
  • [COMMAND_EXECUTION]: The skill executes several shell commands within the CI/CD runner, including the strix CLI tool, git for repository management, jq for processing JSON metadata, and curl for API communication.
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface as it is designed to ingest and analyze untrusted code diffs from pull requests. 1. Ingestion points: Code changes in repository pull requests processed by the strix tool. 2. Boundary markers: No explicit markers or delimiters are defined in the skill instructions to separate code data from agent instructions. 3. Capability inventory: The environment permits shell command execution (strix, git, jq, curl) and network access to vendor APIs. 4. Sanitization: The instructions do not specify any sanitization or filtering of the code content before it is processed by the AI-based tool.
Recommendations
  • HIGH: Downloads and executes remote code from: https://strix.ai/install - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 7, 2026, 06:30 PM
Security Audit — agent-trust-hub — ci-security-scanning-with-strix