Security readout for executives and security teams
Plain-English summary
A crafted prompt can bypass a human approval safeguard and cause an AI agent to run arbitrary Python on its host. Successful abuse could compromise the confidentiality, integrity, and availability of that host. Amazon identifies versions before 0.8.5 as vulnerable.
Executive priority
Treat this as an immediate remediation priority for internet-facing or externally influenced agents with Python tooling. The potential impact is full compromise of the agent host, although active exploitation is not documented in the supplied evidence. Prioritize inventory, upgrade, and deployment verification.
Technical view
The python_repl tool improperly handles LLM prompt input when batch keyword arguments are forwarded. An attacker can pass non_interactive_mode through the batch tool, bypassing the consent gate and enabling arbitrary Python execution. The CVSS 4.0 score is 9.2, with network reachability, low complexity, no privileges, and no user interaction, but attack requirements are present.
Likely exposure
Exposure is likely where strands-agents-tools before 0.8.5 is deployed, python_repl is enabled, the batch tool can reach it, and untrusted prompts can influence tool arguments. The supplied evidence does not establish which applications or configurations enable this path by default.
Exploitation context
The source bundle reports no KEV listing, and the cited information provides no evidence of active exploitation. Exploitation requires a crafted prompt and a usable path through the batch tool to python_repl. Public proof-of-concept availability is not established by the supplied sources.
Researcher notes
CVE-2026-78379 is categorized as CWE-1427. The trust-boundary failure is keyword forwarding that changes python_repl into non-interactive operation, defeating human consent. Assess whether application wrappers, tool allowlists, or prompt-routing controls prevent the affected batch-to-python_repl path. No affected-version granularity beyond versions before 0.8.5 is supported here.
Mitigation direction
- Upgrade strands-agents-tools to version 0.8.5 or later.
- Follow the AWS security bulletin and verify the corrected package reaches every deployed environment.
- Until upgraded, restrict untrusted prompt access to agents exposing python_repl or batch tooling.
Validation and detection
- Inventory all direct and transitive installations of strands-agents-tools.
- Confirm every deployed package version is 0.8.5 or later.
- Identify agents where untrusted prompts can access both batch and python_repl tools.
- Review agent and host logs for unexpected Python execution; no source-provided indicators are available.
Public sources used
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
CWE-1427: Exact CWE lookup
Use the exact CWE identifier as the starting point before reviewing related ATT&CK behavior. Open the exact CWE lookup page first, then review the ATT&CK searches from that MITRE weakness context. This is a Glexia lookup hint, not an official ATT&CK mapping.
Open ATT&CK lookupCVE-2026-78379 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
Open ATT&CK lookup- Severity
- Critical
- CVSS
- 9.2 (4.0)
- Known Exploited
- No
- Published
Vector: CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
CNA and ADP enrichment extracted from CVE v5
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
SSVC decision data
CVSS vector scores
2 official scoresWe collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N——AMZNCVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H2.25.9AMZNVulnerability scoring details
Base CVSS 4.0 score
9.2CriticalVector: CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Base CVSS 3.1 score
8.1HighVector: CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
Vulnerability timeline
Timeline events are normalized from CVE metadata, CNA source timelines, ADP timelines, and KEV metadata when present.
- CVE reservedCVE Program
The CVE ID was reserved by the assigning CNA.
- CVE publishedCVE Program
The CVE record was published.
- CVE updatedCVE Program
The CVE record metadata indicates this as the latest update time.
ADP provider summaries
Source materials
- CVE List V5 sourceCVE List V5
- https://pypi.org/project/strands-agents-tools/0.8.5/CVE reference, AMZN · patch, release-notes
- https://aws.amazon.com/security/security-bulletins/2026-089-aws/CVE reference, AMZN · vendor-advisory
Products and packages named in the record
CWE details
CWE links open Glexia weakness intelligence pages with official CWE context, developer remediation guidance, and related CVE mappings.
Improper Neutralization of Input Used for LLM Prompting
Improper Neutralization of Input Used for LLM Prompting represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.