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.
npx skills add XuanRanL/loamwright-SEO-Skill --skill image-post-processorgit clone --depth 1 https://github.com/XuanRanL/loamwright-SEO-SkillWrote 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/skills/xuanranl/loamwright-seo-skill/image-post-processor)<a href="https://agentmods.dev/skills/xuanranl/loamwright-seo-skill/image-post-processor"><img src="https://agentmods.dev/badge/skills/xuanranl/loamwright-seo-skill/image-post-processor.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00051 | $0.00835 |
| Opus 5 | $0.00026 | $0.00417 |
| Sonnet 5 | $0.00010 | $0.00167 |
| Haiku 4.5 | $0.00005 | $0.00084 |
Grade A, and why
image-post-processor scanned grade A with 1 finding 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 4d 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
subprocess.run(["python", "-m", "scripts.image.exif_stripper", str(png_file)]) How it starts
The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Image Post Processor
Transforms raw PNG outputs from the image pipeline (any provider — openclawroot relay or official OpenAI, both return gpt-image-style PNGs) into production-ready WebP variants.
Inputs
projects/{slug}/assets/images/{article-slug}/{custom_id}.png× 4workspace/{task_id}/image_prompts.json(for alt_text_seed, filename_seed)
4-step pipeline (per image)
import subprocess
from pathlib import Path
images_dir = Path("projects/{slug}/assets/images/{article-slug}/")
for png_file in images_dir.glob("*.png"):
slot_id = png_file.stem
# Step 1: Strip EXIF
subprocess.run(["python", "-m", "scripts.image.exif_stripper", str(png_file)])
# Step 2: WebP convert
webp_file = png_file.with_suffix(".webp")
subprocess.run(["python", "-m", "scripts.image.webp_converter",
str(png_file), "-o", str(webp_file), "-q", "85"])
# Step 3: Srcset 3 sizes
is_cover = slot_id == "cover"
args = ["python", "-m", "scripts.image.srcset_generator",
str(webp_file), "-o", str(images_dir)]
if is_cover:
args.append("--cover") # uses 480/1024/2048
subprocess.run(args)
# (Removed 2026-06-10: the old "Step 4: compress to <200KB" --target-kb loop.
# q85 is the single quality knob now — a hard KB cap on 2K/4K sources just
# re-compressed them to ~q75 and undid the quality upgrade.)
# Step 4: SEO filename rename
article_slug = ws.parent.name # or from state
purpose = lookup_purpose_for_slot(slot_id) # from image_prompts.json
new_name = subprocess.run(["python", "-m", "scripts.image.image_seo_filename",
"--slug", article_slug,
"--purpose", purpose,
"--ext", ".webp"],
capture_output=True, text=True).stdout.strip()
webp_file.rename(images_dir / new_name)
Output
projects/{slug}/assets/images/{article-slug}/
├── best-fishing-rods-2026-cover.webp (3840w)
├── best-fishing-rods-2026-cover-480w.webp
├── best-fishing-rods-2026-cover-1024w.webp
├── best-fishing-rods-2026-cover-2048w.webp
├── best-fishing-rods-2026-section-1.webp (1024w)
├── ... (similar for sections 2, 3)
└── image_meta.json (per-image metadata)
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.
- 4d ago First seen · 89 lines · 0 tokens per session scan A fdc6b1e81c32
image-post-processor is a skill published in the GitHub repository XuanRanL/loamwright-SEO-Skill (47 stars, last pushed 21d ago), licensed Apache-2.0. It adds 51 tokens to every session and 835 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
image-seo-audit
Audit every image on a page for SEO, performance, and accessibility — alt text quality, tiered file-size thresholds, WebP/AVIF format adoption, srcset/sizes responsiveness, lazy loading, fetchpriority on the LCP image, and width/height for CLS — producing an optimization list sorted by file-size savings. Triggers on…
landing-page-audit
Audit a landing page across six conversion dimensions — above-fold clarity, trust signals, form friction, message match against the upstream ad or email, page speed, and mobile experience — each scored 1-10, rolled into an overall score benchmarked against industry averages, with the top 5 fixes ranked by expected…
mcp-modernize
Rewrite an imported (absolutely-positioned) Unbounce page variant as clean responsive HTML through Unbounce MCP. Use when the user asks to modernize a page, make a page responsive, get rid of absolute positioning, make a clean-HTML version of an existing Unbounce page, or edit a page that was built in the Unbounce…
mcp-page-authoring
Best practices for authoring an Unbounce landing page through Unbounce MCP — writing the HTML/CSS/JS for createpagefromhtml, createvariantfromhtml, and updatevariantfromhtml. Use when the user asks to create, build, design, generate, or edit an Unbounce landing page or variant, or to personalize page text by URL…
claude-design
Design one-off HTML artifacts (landing, deck, prototype).
pretext
Build creative browser demos with DOM-free text layout.