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 storyAura/lora-tagging-skills --skill prompt-pyramidgit clone --depth 1 https://github.com/storyAura/lora-tagging-skillsWrote 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/storyaura/lora-tagging-skills/prompt-pyramid)<a href="https://agentmods.dev/skills/storyaura/lora-tagging-skills/prompt-pyramid"><img src="https://agentmods.dev/badge/skills/storyaura/lora-tagging-skills/prompt-pyramid/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/skills/storyaura/lora-tagging-skills/prompt-pyramid"><img src="https://agentmods.dev/badge/skills/storyaura/lora-tagging-skills/prompt-pyramid.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 | $0.00210 | $0.03406 |
| Opus 5 | $0.00105 | $0.01703 |
| Sonnet 5 | $0.00042 | $0.00681 |
| Haiku 4.5 | $0.00021 | $0.00341 |
Grade A, and why
prompt-pyramid 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 12d 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 — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Prompt Pyramid — Visual-Weight Reordering for Diffusion Tags
What this skill does
Take a messy, unordered pile of booru-style tags and return one clean comma-separated string, ordered so the model's attention mechanism locks onto identity and color first, then sculpts the head, then dresses the body outward-to-inward and top-to-bottom, and finally anchors the feet. The goal is: the model never has to "look back" to revise a region it already rendered, which is how color bleeding, floating accessories, and extra limbs happen.
Output format: a single line of comma-separated tags, lowercase, spaces instead of underscores, no parenthetical weights, no artist strings, no duplicates. Nothing else in the final output line unless the user explicitly asks for explanation.
Why ordering matters (the core mental model)
Diffusion text encoders assign higher attention weight to tokens that appear earlier in the prompt. This is not a myth — it's the same positional bias you see in any transformer over a sequence. Two practical consequences:
- Whatever comes first gets rendered first and most "solidly." Later tags are interpreted as modifications to an already-formed region. So if a hair accessory comes before the hair color, the model often paints the accessory color onto the hair, or draws hair that doesn't match the stated color.
- The model hates revising. If head features are split across the prompt (one hair tag at the top, another at the bottom), attention to "the head" flickers on and off, and you get the classic failure modes: duplicate hair accessories, a stray arm, a floating ribbon. Grouping all of one region's tags together lets the model render that region once, confidently, and move on.
So the ordering below isn't aesthetic — it's a sequence that minimizes revision.
The six-tier pyramid
Read this top to bottom. That is the order tags must appear in the final string.
Tier 1 — Identity (the skeleton)
1girl / 1boy / 0girl / multiple girls / solo / crowd · character count · the LoRA/character trigger word
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 168 lines · 210 tokens per session scan A 3a8757bb6fc0
prompt-pyramid is a skill published in the GitHub repository storyAura/lora-tagging-skills (3 stars, last pushed 1mo ago), licensed MIT. It adds 210 tokens to every session and 3,406 once invoked, about $0.0011 per session on Opus 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 skills, from other repositories
shortfilm-prompt
Generate cinematic AI shortfilm prompts (works with Seedance 2.0, Xiaoyunque, Sora, Kling, Jimeng, Veo) using the 5-stage structure from Mx-Shell's Zombie Scavenger. Trigger when the user wants transformation sequences, multi-shot narrative shorts, weapon-charge/combat segments, emotional family/pet/farewell…
video-frame-extractor
A video-frame extraction and analysis helper that selects key frames, describes their contents with a vision model, and creates structured prompts for further creation.
kling-prompter
A prompt-writing guide for Kling 3.0, a video-generation model. It covers short videos, story-driven videos with dialogue or sound, and animating an existing reference image.
happyhorse-prompter
A tool for writing short prompts for HappyHorse 1.0, an AI video generator that can create video with sound. Its prompts put the main subject first, specify the camera, and describe audible sound when needed.
seedance-prompter
A prompt-writing guide for Doubao Seedance 2.0, an AI model that generates videos from written instructions.
Prompt craft for ComfyUI generation
Use when writing prompts for ComfyUI image or video models: choosing tags vs literary format, fixing composition and anatomy artifacts, picking samplers per LoRA, and writing short movement prompts for video. Neutral examples only.