strix-fix-findings

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the strix command-line utility to perform security audits and verify that fixes have been correctly applied.
  • Evidence: Commands such as strix -n -t ./ --scan-mode quick ... are used to run local scans.
  • [EXTERNAL_DOWNLOADS]: Communicates with the Strix Cloud API (app.strix.ai) via curl to manage security scans and retrieve vulnerability details.
  • Evidence: Instructions include API calls to endpoints like /api/v1/scans/{scanId}/rerun and /api/v1/vulnerabilities/{vulnerabilityId}/retest for interacting with the vendor's cloud platform.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes security reports from external sources which could contain instructions designed to influence the agent's behavior during remediation or verification.
  • Ingestion points: Reads findings from local files (vulnerabilities/*.md, vulnerabilities.json, findings.sarif) and cloud API responses (technical_analysis, poc_description, poc_script_code).
  • Boundary markers: Not present; the instructions do not specify the use of delimiters or 'ignore' instructions when parsing these external findings.
  • Capability inventory: The agent can execute shell commands (strix, git), perform network requests (curl), and modify source code to apply fixes.
  • Sanitization: There are no explicit instructions to sanitize or validate the content of the vulnerability reports before processing them.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 02:18 PM
Security Audit — agent-trust-hub — strix-fix-findings