zhihu-scraper

Use when changing or auditing zhihu CLI commands, help text, README command examples, or bare zhihu help semantics.

Use when maintaining this repository after v3.0.1-final, deciding scope, rejecting broad rewrites, or preparing a small stabilization change.

Use when changing Markdown output, image downloading, SQLite writes, save contracts, or fetch workflow persistence behavior.

Use when modifying config schema, runtime config loading, Cookie paths, logs, browser profile paths, or local output path resolution.

Use when updating README, README_EN, CONSTITUTION, AGENTS, install docs, command examples, or governance wording.

Use when changing scraper payload parsing, typed scraper contracts, supported Zhihu URL handling, browser fallback behavior, or anti-drift tests.

This skill should be used when the user asks to "write a readme", "create readme", "generate readme", "improve readme", "audit readme", "review readme", "fix readme", "readme best practices", "readme standard", "perfect readme", or mentions README quality, documentation standards, or developer experience documentation.

README quality standards for engaging, accessible, scannable content including problem-solution hooks, plain language (no unexplained jargon), acronym context, paragraph limits (≤5 lines), benefits-focused language, visual hierarchy, and progressive disclosure. Essential for creating effective README files that welcome and guide users.

Apply when writing code that must be hard to misuse: immutability, visibility, contracts, assertions, resource safety, and type-level invariant enforcement. Trigger on mentions of const, encapsulation, access control, defensive coding, RAII, or 'harder to misuse.' Not for security against adversarial input — use the security skill for that. Language-specific idioms are in references/.

Apply when writing comments, docstrings, READMEs, ADRs, or any form of technical documentation. Trigger on mentions of documentation, comments, docstrings, ADRs, or explaining code. Focuses on concise, non-obvious documentation that captures business logic and decisions the code cannot express. Not for API contract documentation — use the api-design skill for that.

Analyses codebase structure to identify unclear module boundaries and suggests architectural improvements that make the code easier to test and easier for agents to navigate. Invoke with /improve-codebase-architecture. Read-only — produces a prioritised findings report, no code changes.

Apply when restructuring existing code without changing its behavior: extracting methods or classes, moving responsibilities, simplifying conditionals, improving names, or breaking dependencies in legacy code. Trigger on mentions of refactoring, code smells, extract method, legacy code, or technical debt. Language-specific idioms are in references/.

Deep, multi-agent codebase quality audit. Spawns one focused agent per quality dimension in parallel, each going deep with evidence-based analysis, then synthesizes into a scored report. Use when you want a thorough assessment of a codebase's strengths, weaknesses, and risks. Also supports --checkpoint mode for layer coherence audits during orchestration.

Run a cross-session coherence audit after merging a layer. Wraps /autoboard:audit with dimension selection and COHERENCE-REPORT processing. Re-invoke at the start of each new layer.

Comprehensive testing and QA workflow covering unit testing, integration testing, E2E testing, browser automation, and quality assurance.