You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

DB-GPT v0.8.0 β€” Unauthenticated datasource credential disclosure (plaintext password leak) β€” huntr PoC

Private/gated proof-of-concept accompanying a huntr vulnerability report. Access is granted only to the huntr triage bot.

Target: eosphoros-ai/DB-GPT @ v0.8.0 (commit a3f21350, latest release).

The datasource API ships stored DB credentials to the client in cleartext. RDBMSDatasourceParameters tags its password field with metadata={"tags":"privacy"} and str(param) honours it (masks the secret) β€” but BaseParameters.to_dict() (asdict) ignores the tag, and DatasourceService._to_query_response returns param.to_dict() as DatasourceQueryResponse.params. The privacy control masks in one path and leaks the plaintext password in the serializer path β€” a genuinely broken defense, not merely "missing auth".

Contents

  • README.md β€” this file.
  • REPORT.md / POC.md β€” the vulnerability report and walkthrough (added separately).
  • poc/fire_031.py β€” self-contained deterministic PoC (differential oracle: str() masks vs to_dict() leaks, on the same secret). Run network-isolated:
    unshare --user --map-root-user --net python3 poc/fire_031.py
    
  • poc/vendor/ β€” the four real v0.8.0 source files the PoC loads/executes, byte-identical to the released PyPI wheel dbgpt==0.8.0 β€” sha256 in poc/PROVENANCE.txt:
    • parameter_utils.py (BaseParameters.__str__ / .to_dict / masking),
    • manager.py (RegisterParameters / polymorphic metaclass),
    • parameter.py (BaseDatasourceParameters.from_persisted_state),
    • base.py (the RDBMSDatasourceParameters class with the password privacy tag).
  • poc/sandbox_031.log β€” a recorded run (differential oracle + server-sink reproduction, all PASS).

Harmless in-memory dataclass ops only; no network egress; the target repository is never built or installed.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support