CVE-2026-56833: PYSEC-2026-3499 - PraisonAI Dynamic Context history and terminal tools read files outside configured storage via path traversal

CVE-2026-56833 · Severity: high · CVSS 7.5 · Published 2026-10-23

PraisonAI's Dynamic Context module manages agent history and terminal logs on disk. A path traversal vulnerability allows lower-privileged users or prompts calling the history and terminal tools to read arbitrary .jsonl and .log files outside the configured storage directory by manipulating run_id and agent_id parameters. An attacker with access to these tools could disclose sensitive data stored elsewhere on the system.

Technologies: praisonai (PyPI). Vendors: PyPI.

Quick answers

Is CVE-2026-56833 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-56833 being exploited in the wild.
How severe is CVE-2026-56833?
High severity, with a CVSS score of 7.5, published 23 October 2026.
Which products and versions are affected?
praisonai (PyPI) ECOSYSTEM:introduced=3.8.1,fixed=4.6.59.
Is there a fix for CVE-2026-56833?
Yes. praisonai (PyPI): upgrade to 4.6.59 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-56833 attacked?
It is attacked with local access, in a software library.
Am I affected by CVE-2026-56833?
Check whether you run praisonai (PyPI) at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

How to fix

  • praisonai (PyPI): upgrade to 4.6.59 or later
    pip install "praisonai>=4.6.59"

Technical details

The vulnerability exists in HistoryStore._get_history_path() and TerminalLogger._get_log_path() in src/praisonai/praisonai/context/history_store.py and src/praisonai/praisonai/context/terminal_logger.py. These functions construct filesystem paths by joining user-controlled run_id and agent_id parameters without validating that the resolved path remains under the configured base_dir. An attacker can pass an absolute path for run_id or use .. sequences in agent_id to escape the base directory and read files like /etc/passwd or other sensitive .jsonl/.log files. The tools history_tail(), history_get(), history_search(), terminal_tail(), terminal_grep(), and terminal_commands() pass caller-supplied run_id and agent_id directly to these path functions without allowlist validation or relative_to() checks. A patch is available in version 4.6.59 and later.

Affected products

  • PraisonAI PraisonAI >= 3.8.1, <= 4.6.58

Identifiers

  • CVE-2026-56833
  • GHSA-22cj-m4wf-fv2c
  • PYSEC-2026-3499

Timeline

  • 2026-06-18: disclosed: Advisory GHSA-22cj-m4wf-fv2c published

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.