Hookify Rule Author
Write precise Claude Code rules for warnings, blocking, and checks
The skill only instructs users to create local Markdown rules and names the fields being matched; it declares no external service, credential, or upload flow. However, rules can block tool operations or stop sessions without requiring user confirmation, auditability, rollback, or false-positive recovery; permanent disabling is deletion, so points are deducted.
The format, fields, events, and workflow are broadly coherent and common pitfalls are documented. However, the optional action field is described but not demonstrated in examples, and no committed tests cover parsing, matching, abnormal input, or runtime failure feedback. Actual behavior therefore remains unverified and the score is reduced.
Trigger phrases, intended tasks, event types, and field ranges are reasonably clear for writing hookify rules. Boundaries for non-fit scenarios, compatibility, complete input constraints, and Chinese-language semantic triggering are not specified, while environment-fit evidence is limited, so points are deducted.
The document uses progressive sections and provides examples, naming conventions, and some limitations. It has version 0.1.0 but no changelog, named maintenance owner, update path, installation/dependency notes, or troubleshooting FAQ; the referenced examples directory is not verified by the supplied material, so points are deducted.
The document plausibly completes the core task of creating, refining, and disabling rules and provides directly reusable frontmatter, patterns, and message examples. There is no reproducible execution evidence, action/parsing semantics remain uncertain, and users must still test for false matches, so the static cap applies and points are deducted.
The SKILL.md is auditable and contains concrete formats and examples, but there is no committed test suite, CI coverage for this skill's key paths, or third-party execution evidence. The supplied repository workflows do not validate this skill, so only limited traceability is justified.
- Obtain explicit user approval before creating block rules, and retain a recoverable disable or rollback path.
- Validate action handling, event fields, and Python-regex behavior in the target Claude Code/hookify version, especially for advanced conditions.
- Do not submit sensitive commands, paths, or code to online regex testers; prefer local tests and inspect false positives.
What it does & when to use it
This skill helps users create and maintain Hookify rules. Rules are Markdown files with YAML frontmatter stored in `.claude/`, where they watch Bash commands, file edits, stop events, or prompts. It covers simple regex rules, multi-condition rules, warning and blocking actions, and rule testing and disabling. It fits Claude Code users who need security reminders, coding-policy checks, or completion checks.
It guides users through choosing an event, writing Python-style regex or multi-condition matches, and creating .claude/hookify.{name}.local.md files. It explains fields for rule names, enablement, events, matching fields, actions, and Markdown trigger messages. It also documents testing patterns with python3, refining rules, and disabling them with enabled: false.
- A developer wants warnings or blocking for commands such as `rm -rf`, `sudo`, or risky permission changes.
- A team wants to flag `console.log` or security-sensitive code during TypeScript edits.
- A maintainer wants reminders when `.env`, credential, or certificate files are edited.
- A project lead wants a checklist for tests, builds, and documentation before the agent stops.
- A Claude Code user wants a deployment checklist triggered by prompt content.
Pros & cons
- Covers Bash, file-edit, stop, and prompt events.
- Documents both simple regex and multi-condition formats.
- Includes workflows for naming, testing, refining, and disabling rules.
- Supports configurable `warn` and `block` actions with Markdown messages.
- It is explicitly based on Claude Code Hookify files and events, so other platforms require adaptation.
- The source provides no standalone test suite or cross-platform validation evidence.
- It provides authoring guidance rather than a standalone rule execution program.
- The installation information does not specify this skill’s plugin name.
How to install
The skill is located at plugins/hookify/skills/writing-rules/SKILL.md in the repository. The README says the collection can be installed through Claude Code using /plugin install {plugin-name}@claude-plugins-official; the source does not identify the specific {plugin-name} for this skill, so a more precise command cannot be verified.
How to use
Ask Claude Code to “create a hookify rule,” “write a hook rule,” “configure hookify,” or “add a hookify rule.” Then create .claude/hookify.{descriptive-name}.local.md in the project root with YAML frontmatter and a trigger message; rules are read on the next tool use. Supported events are bash, file, stop, prompt, and all.