skillsaw.org

What It Is

Skillsaw is an advanced linter for Agent Context. Its core use case is not traditional code style checking, but reviewing instruction files, plugins, skills, commands, agents, hooks, and marketplace structures used by AI Agents. It claims to provide 40+ to 50 rules covering issues such as weak language, tautology, attention dead zones, embedded secrets, contradictions, inconsistent terminology, context budget, and structural metadata.

Core Capabilities

Based on the crawled content, Skillsaw’s strength lies in its deep combination of “content intelligence” and “Agent instruction engineering.” It can automatically recognize formats such as CLAUDE.md, AGENTS.md, Cursor, Copilot, Gemini, and Kiro, and check consistency across files. For example, the inconsistent terminology rule can detect mixed use of terms such as directory/folder or npm/yarn, helping prevent models from dealing with implicit conflicts during reasoning. The CLI supports quick execution via uvx skillsaw, structural fixes with skillsaw fix, and content-quality fixes with skillsaw fix --llm using any LLM. It also supports parallel execution, scoped re-checks, and file-level rollback.

CI, Ecosystem, and Documentation

Skillsaw provides a GitHub Action that can output violations in CI logs. It also offers a separate review action that posts inline PR comments, with comment deduplication and automatic resolve support. The documentation specifically provides a two-workflow setup to avoid granting write permissions to untrusted PR code, and recommends pinning to a commit SHA, showing a strong awareness of CI security. In terms of ecosystem, it supports custom rules, disable patterns, thresholds, scaffolding generation, and HTML/Markdown documentation generation, making it suitable for teams that want to bring AI instruction assets under engineering governance.

Pricing and Open Source Status

The crawled text does not provide pricing, payment methods, or commercial support information. The page shows the project path as stbenjam/skillsaw and provides GitHub Action usage, but does not clearly state a license, so its open-source licensing cannot be determined directly. For deployment, it can run locally via CLI or in GitHub Actions, but no standalone server-side self-hosting instructions were found.

Pros, Cons, and Who It’s For

Its advantages are precise positioning, a rich rule set, mature CI integration, and support for LLM-powered auto-fixes. Its drawbacks are limited information on commercial support, other CI platforms, licensing, and pricing; the quality of LLM fixes also depends on external models. It is suitable for open-source projects, platform teams, DevOps teams, and technical documentation owners who maintain context files for AI coding tools.

Access from China

No information is provided about access from mainland China, mirrors, or payment options, so availability is rated as unknown. If GitHub or PyPI connectivity is unstable, a proxy or internal artifact mirror may be needed. Alternatives to consider include Promptfoo, CodeRabbit, general-purpose Markdown linters, Semgrep, or custom in-house CI rules.