ai-solutions-architecture

Scans content for filler words, hedging, unsubstantiated claims, and redundancy. Produces a line-level issue list with specific replacements or deletions. Use after writing or revising content to tighten prose before shipping. Activates when auditing prose quality, cutting filler, removing noise, or tightening language across one or more MDX content pages.

Evaluates a draft solutions guide page against a 7-criterion quality rubric and produces a scored assessment with specific issues and actionable revision instructions. Use after writing or revising any content page for the solutions architecture guide to determine whether it should ship, be revised, or be rewritten. Activates when reviewing page quality, checking content completeness against the inventory, scoring draft pages, or running the evaluation step in the write-evaluate-revise loop. The rubric covers completeness, accuracy, conciseness, actionability, code quality, cross-linking, and structure.

Evaluates a skill's output quality by running natural language test instructions through the skill, grading the results against expected findings, and producing a scored assessment. Use to validate that a skill produces correct, complete, and actionable output before relying on it in production workflows. Activates when testing skills, grading skill output, validating skill accuracy, or running the eval-iterate loop on SKILL.md prompts.

Evaluates a page or component against UX heuristics for a developer documentation site. Produces scored assessments with specific issues and actionable fixes. Use before implementing visual changes, after building new layouts or components, or when reviewing page structure and styling. Activates when evaluating design decisions, auditing readability, checking dark/light mode, reviewing navigation patterns, or assessing information hierarchy. Covers scannability, visual hierarchy, consistency, responsiveness, dark mode, and pattern appropriateness.

Extracts and inventories educational or technical content from a scraped web page into a standardized markdown inventory record. Use when processing content from any browsed or scraped web page — training courses, documentation, tutorials, blog posts — that needs to be captured with topic tags, importance ratings, code examples, key concepts, patterns, and gotchas for later synthesis into solutions guide pages. Activates during content scraping workflows, inventory building, or when the agent has page content that needs structured capture before writing guide pages.

Synthesizes raw content inventory entries into a polished, practical MDX solutions guide page with runnable code examples, callouts, diagrams, and cross-references. Use when creating or rewriting a solutions guide page from content inventory entries that have been mapped to a specific topic or site section. Activates during the page writing phase of the content pipeline, after content extraction and mapping are complete. Produces pages that follow the project's page template, writing voice, and file conventions. After writing, the content-evaluation skill should be used to score the result.