Skills
Navigate
skills/zxkane/aws-skills/aws-mcp-setup/Gen Agent Trust Hub

aws-mcp-setup

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the mcp-proxy-for-aws package and the uv package manager from official package registries.
  • [REMOTE_CODE_EXECUTION]: Uses the uvx tool to execute the mcp-proxy-for-aws package from a remote repository, which is a standard procedure for running the latest AWS MCP proxy tools.
  • [COMMAND_EXECUTION]: Executes system and tool-specific commands to verify the environment and identity, including aws sts get-caller-identity, claude mcp, and which.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and processes the contents of configuration files which could technically contain untrusted data, though the risk is low as it specifically filters for AWS-related configuration keys.
  • Ingestion points: Reads from .mcp.json and ~/.claude.json via cat commands (SKILL.md).
  • Boundary markers: Absent; the skill relies on grep to filter for specific configuration keys.
  • Capability inventory: Includes uvx execution, pip install, and AWS identity verification (SKILL.md).
  • Sanitization: Absent; configuration data is processed directly via shell pipelines.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 08:29 AM