whats-new

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the official Claude Code changelog from the Anthropic organization's GitHub repository.
  • [COMMAND_EXECUTION]: Uses Bash to check the local installation version and maintains a version tracking file at ~/.claude/whats-new-in-claude-last-version.
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes external content from a GitHub repository and search results from the web.
  • Ingestion points: External data enters the context through WebFetch of the CHANGELOG.md file and via research agents searching anthropic.com and code.claude.com.
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are provided to the research agents.
  • Capability inventory: The skill utilizes Bash (local file writing), WebFetch (network access), and Task (agent orchestration).
  • Sanitization: There is no evidence of sanitization or filtering of the retrieved markdown content before it is passed to sub-agents or displayed to the user.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 11:07 PM