Skillsaw

md files. The tool addresses issues such as weak language, contradictions, attention dead zones, and structural inconsistencies in these files, with the aim of improving the quality and reliability of agent context and behavior.

Skillsaw offers more than 40 research-backed rules to detect problems like tautological instructions, embedded secrets, and other content flaws. It provides deterministic autofixes through the 'skillsaw fix' command, and supports the use of a Skills plugin to assist coding agents in applying fixes. The platform is extensible, allowing users to define custom rules, install rule plugins via pip, and adjust per-rule thresholds to tailor the tool to specific project needs.

Integration with continuous integration systems such as GitHub and GitLab is supported, enabling inline pull request comments, deduplication, and automatic thread resolution. Skillsaw also includes scaffolding features, with the 'skillsaw add' command generating plugins, skills, commands, agents, and hooks that follow best-practice structures. For documentation, the tool can generate HTML or Markdown files for plugins and marketplaces using 'skillsaw docs'.

Skillsaw is delivered as a command-line interface (CLI) and is intended for developers working with AI agent skills and plugins.