Disclosed vulnerability
AWS Kiro: arbitrary command execution via indirect prompt injection writing agent config files
AIR-2025-0116 AIR record · 2025-08-aws-kiro-rce
From the supporting source
AWS Kiro was vulnerable to arbitrary command execution via indirect prompt injection.embracethered.com
The record
Disclosed
2025-08-26
Vendor
Amazon Web Services
Product
AWS Kiro IDE
Risk domain
Access control and permissions › Privilege escalation
Also touches
Confused deputy
System type
Generative model
Who acted
Agent acted — Kiro wrote config files and executed arbitrary commands
Attack surface
Coding agent
Autonomy level
Semi-autonomous
Initial vector
Indirect prompt injection source code
Impact type
Remote code execution, Code execution
Confidence
High
Submitted by
EAI Internal
Crosswalk
- OWASP Agentic Top 10 (2026)
- ASI03 – Identity and Privilege Abuse
- MITRE ATLAS
- AML.T0055 – Unsecured Credentials
- NIST AI RMF
- GOVERN 6.1
- EU AI Act
- Article 13
- ISO/IEC
- TR 24028; 42001; 23894
- AIUC-1
- B. Security
- Scope
- tool misuse
- Mitigation
- Apply least-privilege principle; Regular policy audits.
- Taxonomy
- Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
Kiro developer workstations before v0.1.42
Capability abused
Defense & remediation
Recorded defense
command approval prompts, bypassed by writing to .vscode/settings.json or .kiro/settings/mcp.json
Guardrail kinds
Human approval
Guardrail outcome
Silent
Fix shipped
fixed 2025-08-05 in v0.1.42; AWS issued no CVE and no public advisory
Taxonomy & references
Curator notes
Reported 2025-07-15, the day Kiro launched. Also reported as: 'AWS Kiro: arbitrary code execution via indirect prompt injection'.
Supporting sources
Cite as
Enkrypt AI. "AWS Kiro: arbitrary command execution via indirect prompt injection writing agent config files". AIR, AIR-2025-0116. https://agentincident.vercel.app/i/AIR-2025-0116/ (accessed 2026-10-19).