awsflow-s3

Warn

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: MEDIUMINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from S3 objects, creating a surface for indirect instructions to influence agent behavior.
  • Ingestion points: The GetObject command (with AsText: true) and SelectObjectContent read external data into the agent context (SKILL.md).
  • Boundary markers: Absent. The skill lacks delimiters or instructions to treat object data as untrusted content.
  • Capability inventory: The agent has extensive permissions including PutObject, DeleteObject, PutBucketPolicy, and DeleteBucketPolicy (SKILL.md).
  • Sanitization: Absent. There is no validation or filtering of content retrieved from S3.- [PRIVILEGE_ESCALATION]: The PutBucketPolicy and DeleteBucketPolicy commands allow the agent to modify or remove security constraints on S3 buckets, which could lead to unauthorized access or public data exposure.- [DATA_EXFILTRATION]: The GetObject command facilitates the reading of object content (AsText) or downloading to temporary storage (DownloadToTemp), providing a path for sensitive data to be extracted from secure storage.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 5, 2026, 11:07 AM
Security Audit — agent-trust-hub — awsflow-s3