s3-upload

Pass

Audited by Gen Agent Trust Hub on Aug 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data (object keys and metadata) from external S3-compatible storage.
  • Ingestion points: The list, find, and head commands in scripts/s3-cli.js retrieve object keys and metadata from remote S3-compatible servers and output them to the agent.
  • Boundary markers: Although the CLI provides structured JSON output (via the --json flag) which helps delineate data, it lacks explicit boundary markers or instructions to the agent to disregard instructions embedded within the retrieved object metadata.
  • Capability inventory: The skill possesses significant capabilities, including writing files to the local system (download), deleting remote objects (delete), and making network requests to arbitrary S3-compatible endpoints defined in environment variables.
  • Sanitization: The script does not perform sanitization or validation of object keys or metadata before presenting them to the agent, allowing potentially malicious strings to enter the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 22, 2026, 08:51 AM
Security Audit — agent-trust-hub — s3-upload