| name | claude-anthropic |
| description | Govern Claude/Anthropic configuration of the Vue SSR Starter Kit (e-xode/vue-ssr): rules and audit method for CLAUDE.md, skills under .claude/skills/, path-scoped rules under .claude/rules/, sub-agents under .claude/agents/, and hooks under .claude/hooks/. Trigger when creating, modifying, reviewing or auditing a skill / rule / sub-agent / CLAUDE.md / hook, when deciding between a rule vs a skill for a constraint, or when asking about Anthropic doctrine (model spec, agent design, progressive disclosure). Co-load with skill-creator when authoring a skill. Don't use for: the create/eval/iterate loop itself (→ skill-creator), post-task code validation (→ vue-ssr-hooks + hooks agent), app architecture (→ vue-ssr-architecture), or auth flow (→ vue-ssr-auth). |
Claude / Anthropic configuration — Vue SSR Starter Kit
Owns the rules and audit method for the project's .claude/ folder and CLAUDE.md. Source of doctrine for how this project applies Anthropic's skill/agent/hook patterns.
What this skill does (and does not)
| In scope | Out of scope |
|---|
Rules for writing/editing CLAUDE.md | Validating modified application code (→ vue-ssr-hooks + hooks agent) |
| Anatomy of a project skill (frontmatter, references, budget) | The workflow of drafting and evaluating a new skill (→ skill-creator) |
| Anatomy of a project sub-agent | App architecture knowledge (→ vue-ssr-architecture) |
Anatomy of path-scoped rules (.claude/rules/) | Auth flow and security (→ vue-ssr-auth) |
| Hook events, format, project status | Deployment and CI/CD (→ vue-ssr-deployment) |
Audit checklist + automated scripts/audit.py | |
| Anthropic doctrine: progressive disclosure, agent design | |
Division of responsibilities — claude-anthropic ↔ skill-creator
| Concern | Owner |
|---|
| Generic create/evaluate/iterate workflow | skill-creator |
| Eval harness, test prompts, assertions | skill-creator |
| Description-optimisation guidelines | skill-creator |
| Project description conventions (discriminating, pushy, anti-triggers) | claude-anthropic |
Skill naming, placement under .claude/skills/, folder layout | claude-anthropic |
SKILL.md anatomy and token budget for this project | claude-anthropic |
| Project anti-patterns | claude-anthropic |
Post-creation audit (scripts/audit.py) + CLAUDE.md index update | claude-anthropic |
| Anthropic doctrine (model spec, progressive disclosure, agent design) | claude-anthropic |
When this skill hands off, use the convention: ➜ See skill: skill-creator — <reason>.
Core rules (the strict minimum)
CLAUDE.md ≤ 10 KB / ~2500 tokens. Hard rules only. Knowledge belongs in skills. See references/claude-md-anatomy.md.
- One skill = one subject. Discriminating description, explicit anti-triggers, pushy phrasing to combat under-triggering. See references/skill-anatomy.md.
SKILL.md is method + index. Detailed knowledge goes in references/. Keep SKILL.md under ~500 lines (~50 KB hard warn).
- Sub-agents have a strict contract: clear scope, no validation (except the
hooks agent), structured return. See references/agent-anatomy.md.
- No duplication between skills. Search existing skills before authoring a new one. If overlap is unavoidable, define a "Division of responsibilities" table in both skills.
- English only in all persisted artefacts (skills, references, agents, CLAUDE.md, commits, PRs).
- No code comments in
CLAUDE.md or SKILL.md outside fenced code blocks. Prose is the medium.
- Cross-references use a stable convention:
➜ See skill: <name> — <reason> (greppable, visible in diffs).
name in frontmatter == folder name. Mechanical, enforced by scripts/audit.py.
- Scripts belong to a skill. Executable tooling lives in
.claude/skills/<owner>/scripts/, never in a top-level .claude/scripts/ pool. Enforced by scripts/audit.py.
- Description budget: 1,536 characters. Put the most important trigger information first.
- Rules are lightweight guardrails.
.claude/rules/ files are path-scoped constraints (< 2 KB, imperative, no references). Use rules for hard DON'Ts tied to specific file paths; use skills for how-to procedures. See references/rules-anatomy.md.
Audit method
Run before any non-trivial change to .claude/ and after creating/modifying a skill.
Step 1 — Automated checks
python3 .claude/skills/claude-anthropic/scripts/audit.py
python3 .claude/skills/claude-anthropic/scripts/audit.py --json
Covers 14 mechanical checks (file sizes, frontmatter validity, name↔folder, description length + anti-trigger presence, duplicate names, broken relative links, agent frontmatter, agent↔CLAUDE.md cross-refs, skill-index↔folder coherence, English-only heuristic, no code comments in markdown, no global scripts pool, rules structure validation). Exit code 1 on any error.
Step 2 — Manual checks
Walk references/audit-checklist.md. Each item is tagged [AUTO] or [MANUAL].
Step 3 — Anti-pattern sweep
Cross-check against references/antipatterns.md.
Step 4 — Propose, never auto-fix
If issues are found, propose corrections to the user. Never silently rewrite a skill, an agent, or CLAUDE.md.
Workflows
Create a new skill
- (this skill) Confirm no existing skill covers the topic. Define scope, name (kebab-case, prefix
vue-ssr- if project-scoped), and the discriminating description.
- (this skill) Plan the
SKILL.md structure and references/ split.
- ➜ See skill: skill-creator — drives the draft, test prompts, eval iteration.
- (this skill) Run
scripts/audit.py and the manual checklist.
- (this skill) Append the skill to the Skills index in
CLAUDE.md.
- (this skill) If twin skills, add "Division of responsibilities" table in both.
Modify CLAUDE.md
- Confirm the change is a hard rule (operational, project-wide, non-negotiable).
- Edit with smallest surface; preserve section structure.
- Re-run
scripts/audit.py — enforces 10 KB budget and cross-refs.
Add a sub-agent
- Read references/agent-anatomy.md.
- Create
.claude/agents/<name>.md with frontmatter (name, description, tools, optional model).
- Add a row to
## Agents directory in CLAUDE.md.
- Run
scripts/audit.py.
Add or modify a rule
- Read references/rules-anatomy.md.
- Confirm content is a constraint (not knowledge). If > 2 KB, it should be a skill.
- Create
.claude/rules/<descriptive-name>.md with paths: frontmatter.
- Run
scripts/audit.py.
Where to look (routing table)