The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Python subprocess module to execute git commands for branch discovery, diff generation, and commit history retrieval. It also calls the diff2html command-line utility. All calls use argument lists rather than shell strings, mitigating shell injection risks.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of diff2html-cli via npm. This is a well-known open-source utility for formatting git outputs. The viewer also loads CSS and JavaScript from the Cloudflare CDN, a trusted service.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from git diffs, which may include untrusted data if the repository contains contributions from external parties. The agent is instructed to read these diffs and generate summaries, which creates a surface where malicious instructions in code comments could potentially influence the agent's behavior. However, the risk is limited to the summary generation context. Ingestion points: Git diff content read via git diff and git show commands in diff-view.py. Boundary markers: None explicitly defined in the prompt instructions to isolate untrusted diff data. Capability inventory: Subprocess execution for git operations and local HTTP server hosting. Sanitization: Employs escapeHtml in the frontend JavaScript to prevent cross-site scripting (XSS) when rendering file paths and summaries in the browser.

git-diff-browser