AgentLogic is a rule, skill, and workflow governance framework for coding agents. It includes defensive security auditing, clean-room reverse-engineering, and authorized CTF/sandbox orchestration skills (skills/reverse-skill/, decompile-reconstruction) intended strictly for:
- Authorized security assessments, vulnerability research, and defensive hardening
- Clean-room interoperability analysis and authorized binary/source auditing
- Sandboxed educational and CTF environments
These capabilities MUST NOT be used for unauthorized access, exploitation of unowned systems, malware authoring, or bypassing access controls.
If you discover a security vulnerability, accidental secret, or unsafe rule/script behavior in this repository:
- Do not open a public issue with sensitive details or exploit payloads.
- Report it privately via GitHub Security Advisories (once public) or directly to the repository maintainer (
UIhoshi). - Include affected file paths, reproduction steps, and suggested remediation where possible.
- Personal paths, credentials, API tokens, and private
reference/archives must never be committed. - Security/pentest reference documentation must defang active script/payload tags to prevent accidental execution or scanner false-positives.