The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/hotspots.py and its test suite scripts/test_hotspots.py use the subprocess.run method to execute git commands (e.g., git log, git ls-files, git diff) and uv for script execution. The commands are constructed using list-based arguments without the shell=True parameter, which is a standard security practice that prevents shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes source code and comments from the local repository. There is an inherent attack surface where malicious instructions or deceptive metrics could be embedded in the code being analyzed to influence the agent's refactoring recommendations.
Ingestion points: The hotspots.py script reads source file content from the local repository via the read() function.
Boundary markers: The instructions do not define explicit delimiters when the agent reads the identified 'hotspots' in the source code, though it is guided by the quantitative metrics output from the script.
Capability inventory: The skill has the capability to execute shell commands via git and uv. The instructions also mention using secondary skills (bento-refactor, bento-ship) to modify code and create pull requests.
Sanitization: The skill performs code analysis using the lizard library and Python's ast module, but does not sanitize the raw source code content before the AI agent reviews it.
[EXTERNAL_DOWNLOADS]: The skill requires the lizard library for code complexity analysis. This dependency is declared in a PEP 723 script block within scripts/hotspots.py. lizard is a well-known, established open-source tool used for multi-language software metrics.

bento-hotspots