The Agent Skills Directory

[Command Execution]: The skill utilizes local shell commands, including git, find, and various build/test tools like npm, pytest, and make. These commands are used to discover code changes and validate the integrity of the project after simplification. This behavior is expected for the skill's primary purpose but assumes the execution environment is appropriately scoped.
[Indirect Prompt Injection Surface]: The skill processes and analyzes recently modified source code files. These files act as untrusted external data sources that could potentially contain embedded instructions designed to influence the agent's simplification logic or PR generation.
Ingestion points: Reads contents of files identified in Phase 1 and 2 (e.g., .go, .js, .py).
Boundary markers: The skill does not explicitly define delimiters or instructions to ignore embedded commands within the source files.
Capability inventory: Includes the ability to write file changes, execute shell commands for testing/building, and create pull requests.
Sanitization: The instructions do not specify sanitization or filtering of the code content before analysis.

code-simplifier