signing-in-to-aws

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructs the agent to execute precondition checks (aws --version, aws sts get-caller-identity, aws configure list) "silently" before asking for user confirmation. This behavior bypasses explicit user consent for initial system discovery commands.
  • [COMMAND_EXECUTION]: The skill uses local shell execution for all its primary functions, including aws login, aws sts get-caller-identity, and aws configure list to manage and verify authentication states.
  • [DATA_EXFILTRATION]: The skill accesses sensitive AWS environment metadata, specifically the AWS Account ID and ARN via aws sts get-caller-identity. While the skill does not attempt to send this data to an external URL, the access happens without initial user approval.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect injection surface by processing untrusted data from the local environment.
  • Ingestion points: The skill reads and processes the output of shell commands like aws --version and aws sts get-caller-identity within SKILL.md.
  • Boundary markers: None identified; command output is treated as trusted context for the agent's flow logic.
  • Capability inventory: The agent has full access to the local shell to run the aws CLI and potentially other commands requested by the user.
  • Sanitization: There are no instructions or mechanisms mentioned to sanitize or validate the output from the shell commands before the agent acts upon it.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:55 AM
Security Audit — agent-trust-hub — signing-in-to-aws