The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses WebFetch to retrieve the changelog from https://raw.githubusercontent.com/anthropics/claude-code/main/CHANGELOG.md. This is a trusted source belonging to a well-known organization (Anthropic). The content is fetched for analysis purposes and is not executed.
[COMMAND_EXECUTION]: The skill utilizes basic Bash commands like cat and jq to read and parse the local state file .claude-code-version-check.json. These operations are limited to local configuration management and do not pose a security risk.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data (the changelog). While this constitutes an attack surface, the source is a trusted repository, and the skill's purpose is to summarize and categorize the information for a human reviewer. This is considered low risk and within standard agent operational parameters.

changelog-review