update-buttercut

Pass

Audited by Gen Agent Trust Hub on Oct 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes several shell commands to manage the update process, including git rev-parse, git stash, git checkout, git pull, bundle install, and ruby. It uses GIT_TERMINAL_PROMPT=0 to ensure non-interactive git operations.\n- [EXTERNAL_DOWNLOADS]: The skill performs git pull origin main to download new source code from the remote repository associated with the origin alias.\n- [REMOTE_CODE_EXECUTION]: The skill runs bundle install, which downloads and installs Ruby gems from external registries. These packages may contain scripts that execute code during the installation process.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and summarize the CHANGELOG.md file after updating. This creates an attack surface for indirect prompt injection if the source repository were compromised.\n
  • Ingestion points: Content from CHANGELOG.md is read via git diff into the agent's context.\n
  • Boundary markers: No specific delimiters or safety warnings are provided to the agent to treat this content as untrusted data.\n
  • Capability inventory: The agent has permissions to execute shell commands (git, bundle, ruby) and perform network operations.\n
  • Sanitization: No sanitization or filtering of the changelog text is performed.\n- [DYNAMIC_EXECUTION]: The skill executes a local Ruby script (lib/buttercut/library.rb) to identify the software edition, running code that belongs to the application being updated.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 4, 2026, 03:05 PM