The Agent Skills Directory

[SAFE]: The skill is designed for improving code quality and performs standard development actions such as file system analysis, code modification, and running test suites.
[COMMAND_EXECUTION]: Uses Bash to execute project-specific test suites (npm test, pytest, go test) and manage local scratch directories. These actions are within the expected behavior for a refactoring tool.
[DATA_EXFILTRATION]: While the skill reads local source code and project specifications, no network activity or external data transmission patterns were identified.
[PROMPT_INJECTION]: The skill ingests local code into a subagent prompt, presenting a surface for indirect prompt injection. However, risk is mitigated by the local nature of the data, the requirement for user approval before execution, and the validation of all changes against existing tests.
Ingestion points: Local codebase files and directories specified by the user (SKILL.md Steps 1 and 3).
Boundary markers: No specific delimiters or safety instructions are used to encapsulate source code when passed to subagents.
Capability inventory: Access to file system tools (Read, Write, Edit), shell execution (Bash), and subagent creation (Agent).
Sanitization: The skill does not perform sanitization or filtering of the source code content before analysis or processing.

quality-refactor