skills/martinffx/atelier/spec-install/Gen Agent Trust Hub

spec-install

Warn

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill leverages npx and bunx to pull and execute arbitrary code from the NPM registry. Specifically, the skills and @every-env/compound-plugin packages are executed to perform installation and synchronization tasks.
  • [DATA_EXFILTRATION]: The command bunx @every-env/compound-plugin sync is used to read personal configuration data, such as ~/.claude/settings.json. This file often contains sensitive information, including API keys or session tokens. Transferring this data via an external, third-party package represents a high risk of exfiltration.
  • [COMMAND_EXECUTION]: The skill performs extensive file system operations, including creating directories and symlinks within the user's home directory configuration folders (e.g., ~/.config/opencode/ and ~/.claude/).
  • [EXTERNAL_DOWNLOADS]: Downloads and installs external components from the public NPM registry. These dependencies are not version-locked, allowing for potential supply chain attacks if the packages are compromised.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 12, 2026, 08:00 AM
Security Audit — agent-trust-hub — spec-install