framework-update

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The script update-from-upstream.mjs downloads an update engine script from a remote Git repository and executes it locally using Node.js. This allows the skill to utilize the latest merging logic from the framework source.- [COMMAND_EXECUTION]: The skill uses the git command-line interface to perform repository management tasks, such as cloning, fetching, and merging code. It also executes Node.js commands to run internal scripts and validate JSON files.- [EXTERNAL_DOWNLOADS]: The skill performs network requests to Git remotes to fetch framework updates. By default, it connects to the author's official repository at github.com/duc01226/easy-claude.- [DYNAMIC_EXECUTION]: The update process involves reading a script from the remote repository and writing it to a temporary file (tmp/framework-update/engine/framework-update.mjs) which is then executed as a separate Node.js process.- [INDIRECT_PROMPT_INJECTION]: As the skill merges code from external sources, it is susceptible to processing malicious instructions if the upstream repository were compromised. \n
  • Ingestion points: Remote Git repositories defined in the skill's parameters or the project lock file.\n
  • Boundary markers: None; the skill relies on the agent and user to review and resolve conflicts.\n
  • Capability inventory: The skill can execute system commands (spawnSync) and modify project files.\n
  • Sanitization: Relies on manual inspection during the merge resolution process as documented in the provided merge guide.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 06:42 PM