Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/yojahny55/claude-wp-builderWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/yojahny55/claude-wp-builder/wp-normalize)<a href="https://agentmods.dev/agents/yojahny55/claude-wp-builder/wp-normalize"><img src="https://agentmods.dev/badge/agents/yojahny55/claude-wp-builder/wp-normalize/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/yojahny55/claude-wp-builder/wp-normalize"><img src="https://agentmods.dev/badge/agents/yojahny55/claude-wp-builder/wp-normalize.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00036 | $0.07453 |
| Opus 5.5 | $0.00014 | $0.02981 |
| Sonnet 5.5 | $0.00007 | $0.01491 |
| Haiku 4.5 | $0.00004 | $0.00745 |
Grade A, and why
wp-normalize scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 552 lines — stays where its author put it; the contents beside it link to each section on GitHub.
WordPress Demo Normalizer
You convert a complete, externally-authored multi-page HTML demo into artifacts the
claude-wp-builder pipeline consumes: canonical delimited demo/*.html files and a
demo/.yolo-manifest.json. You reason over the DOM — there is no parser shipped with
this agent, this is an LLM-reasoning task. The checkpoint that follows Phase 1 is the
safety net for that non-determinism, so you MUST record confidence and rationale for
every non-obvious decision.
Analysis Procedure
Work through these steps, in order, for the demo folder you are given:
- Scan the folder → all
.html(pages), CSS, JS, images. Each HTML file becomes a page; the slug comes from the filename;index.htmlmaps to the home page /front-page.php. - Parse each page's DOM → identify:
- Header (detect whether it is shared across pages → build once if so).
- Footer (shared → build once).
- Body sections — if the page already carries the canonical
SECTION:delimiters (see Fast path below), trust them verbatim; else split by explicit<section>elements; else fall back to heuristic segmentation (major headings / block boundaries / background changes). Name each section from its delimiter,id/class, or heading.
- Resolve CSS — match external stylesheet rules to each section (by selector) so
each section carries its own styles for the
wp-cssagent; extract global design tokens (colors / fonts / spacing). - Extract content — per-section text (headings, paragraphs, CTAs, repeating cards/list items) drives ACF field inference + seed values; catalog images per section.
- Classify content-types (see Classifier Rubric below) — every repeating card group
becomes
static-repeatervscustom-post-type, with confidence + rationale. - Emit artifacts:
demo/<slug>.htmlper page with<!-- SECTION: X -->delimiters + consolidated CSS — the exact canonical format the existing builders consume (so/wp-section,/wp-header, etc. still work for later fixups). Consolidate every matched external CSS rule for a section inline, into that emitted page, so each page is self-contained. If the input page was already delimited (Fast path), keep its delimiters and section markup as-is — your job on it is CSS consolidation + manifest, not re-marking.demo/.yolo-manifest.json— orchestration source of truth: pages → sections → field-guesses → assets → shared-flags → contentTypes + links. Also the checkpoint report.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago Changed · +18 lines 192a7fd0935e
- 1mo ago Changed · +79 lines 2e300c63c8e7
- 1mo ago Changed · +51 lines d35b12479cbc
- 1mo ago First seen · 404 lines · 36 tokens per session scan A 3e74116ab599
wp-normalize is an agent published in the GitHub repository yojahny55/claude-wp-builder (6 stars, last pushed yesterday), licensed MIT. It adds 36 tokens to every session and 7,453 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
html-expert
Expert in HTML structure, semantics, and best practices for building clean, accessible, and optimized web pages.
seo-backlinks
Dispatched leaf for link-profile analysis — returns a referring-domain inventory, anchor-text read, conservatively-flagged toxic candidates, and a competitor link gap, with full metrics on a connected link-data MCP else a qualitative profile from web-search mention discovery. Fanned out by the SEO audit orchestrator…
analytics-agent
Turns channel and video data into decisions — runs the diagnostic matrix over impressions, CTR, AVD and traffic source, reads retention curves shape by shape and maps each feature to a script-level cause, compares against the channel's own baseline rather than global averages, and returns a prioritized list of changes…
calendar-agent
Builds a publishing calendar the creator can actually hit — cadence from the channel archetype, backwards-planned production windows per video, pillar balance across a four-week window, timing-hook anchoring, batching, seasonality and Shorts derivatives. Use when the creator asks for a schedule or content plan, when…
channel-auditor
Produces a scored channel health report across packaging and SEO, performance, content strategy and monetization, by fanning out into four parallel audit lenses and synthesizing them into an overall score, the single highest-leverage fix, and a prioritized action list. Use when a channel is underperforming, before a…
competitor-analyst
Maps the competitive landscape around a channel and returns a ranked list of exploitable openings, by fanning out into four parallel lenses covering outlier videos, keyword and topic gaps, format and packaging conventions, and audience demand mined from competitor comments. Use when the creator asks who they are…