defense-researcher
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides installation instructions that reference its own source code on GitHub (github.com/theneoai/awesome-skills). These URLs point to the official repository of the skill's author for deployment on platforms like OpenCode, OpenClaw, and Claude Code. This is standard functionality for skill distribution and does not involve untrusted third-party code.
- [DATA_EXFILTRATION]: The skill mentions local file paths such as '
/.opencode/skills/' and '/.claude/CLAUDE.md' within its platform support documentation. These references describe the standard installation and configuration directories for the respective AI agent platforms and do not indicate unauthorized file access or exfiltration. - [PROMPT_INJECTION]: The skill includes a robust 'Ethical Boundary' and 'Scenario 3: Anti-Pattern' section which explicitly instructs the agent to deny requests for weapons development or offensive capability design. This is an embedded safety guardrail rather than a malicious injection.
- [COMMAND_EXECUTION]: No shell commands, script execution, or system-level operations are initiated by the skill. It is a purely instructional document for the AI agent.
Audit Metadata