devcontainer-setup

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Grants the container NET_ADMIN capability to support network isolation via iptables and ipset.
  • [EXTERNAL_DOWNLOADS]: Fetches official devcontainer features from the GitHub Container Registry.
  • [EXTERNAL_DOWNLOADS]: Installs Claude Code plugins from trusted organizations (Anthropic) and reputable security vendors (Trail of Bits).
  • [COMMAND_EXECUTION]: Generates and executes post-installation shell and Python scripts (post_install.py, install.sh) to automate environment configuration.
  • [PROMPT_INJECTION]: The skill processes project-controlled metadata which presents an indirect prompt injection surface.
  • Ingestion points: Reads metadata from package.json, pyproject.toml, Cargo.toml, and go.mod.
  • Boundary markers: Absent; the skill does not wrap untrusted project data in delimiters.
  • Capability inventory: The skill can write and execute Dockerfiles, Python scripts, and shell scripts.
  • Sanitization: Absent; project-sourced strings (like project names) are interpolated into scripts and configurations without explicit validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 04:34 PM
Security Audit — agent-trust-hub — devcontainer-setup