agent-code-analyzer

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS] (MEDIUM): The skill repeatedly calls npx claude-flow@alpha in its hooks and workflow phases. This package is not from a trusted organization (e.g., Anthropics, Google, Microsoft) and is pulled from the public npm registry at runtime, which poses a supply chain risk.
  • [REMOTE_CODE_EXECUTION] (MEDIUM): The use of npx to fetch and execute a remote script/package at runtime without pinned versions or integrity hashes constitutes remote code execution of unverified content.
  • [COMMAND_EXECUTION] (MEDIUM): The skill executes multiple shell commands to perform memory operations and task hooks. There is a potential for command injection in Phase 3 where ${results} (the output of code analysis) is interpolated directly into a shell command line: npx ... --value "${results}". If the analyzed code contains characters that break the shell string, it could lead to arbitrary command execution.
  • [PROMPT_INJECTION] (LOW): The skill is highly susceptible to indirect prompt injection (Category 8).
  • Ingestion points: Code files processed during static analysis and deep analysis phases (SKILL.md).
  • Boundary markers: None present. The skill does not define delimiters or instructions to ignore embedded commands in the code being analyzed.
  • Capability inventory: Subprocess execution (npx), memory read/write, and notification hooks.
  • Sanitization: Absent. There is no evidence that the results of the code analysis are sanitized before being stored in memory or sent via notifications.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:10 PM