Agent skill

Anti Slop Image Direction

by BlackBeltTechnology in BlackBeltTechnology/pi-agent-dashboard

Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for…

MITAuto-check: notesWriting & Content

Install Anti Slop Image Direction

skills CLI
$ npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-image-direction -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install BlackBeltTechnology/pi-agent-dashboard anti-slop-image-direction --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/BlackBeltTechnology/pi-agent-dashboard.git skills-src && mkdir -p .claude/skills && cp -r skills-src/packages/anti-slop/.pi/skills/anti-slop-image-direction .claude/skills/anti-slop-image-direction && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
anti-slop-image-direction
GitHub stars
315
Token cost
~2.1k tokens
SKILL.md length
1,058 words
Files
1
Skills in repo
70
Repo updated
First seen
Licence
MIT

At a glance

Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for…

  • Works in 4 steps: Plan, then confirm → Generate → Analyze → refs/direction.md → …
  • Tasks that involve Humanizing AI text
  • SKILL.md covers When it applies, Images are experimental; the…, Procedure and Direction, not source, plus 4 more sections
  • Calls npx; needs GEMINI_API_KEY

What it does

Anti Slop Image Direction is an agent skill from BlackBeltTechnology/pi-agent-dashboard. Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for the mockup loop. Images are direction, never source. Never for product UI. Triggers: "generate design references", "image direction for this page", "show me visual directions".

Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Writing & Content, covering Humanizing AI text and Image generation. It works with Google Gemini and npm. The repository describes itself as: Real-time web dashboard for pi coding-agent sessions. Multi-session view, live chat mirroring, integrated terminal, diff viewer, pi-flows execution, and mobile-first remote… The licence is MIT.

When your agent uses it

  • Tasks that involve Humanizing AI text
  • Tasks that involve Image generation

Example prompts

  • “generate design references”
  • “image direction for this page”
  • “show me visual directions”
  • “/anti-slop-image-direction”

Requirements

  • Node.js
  • A credential in GEMINI_API_KEY

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Plan, then confirm
  2. Generate
  3. Analyze → refs/direction.md
  4. Hand-off

What it can do on your machine

Read from SKILL.md and the folder at commit 7a2d171. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npx

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npx, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • GEMINI_API_KEY

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Anti Slop Image Direction loads about 2.1k tokens when it runs. Until then it costs about 100 tokens; SKILL.md has 1,058 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~100
When it runs · the whole SKILL.md, loaded when a task matches
~2.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:57
    `.env`) or `pi` (`--backend pi`, an OpenRouter credential inside pi; a

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from BlackBeltTechnology/pi-agent-dashboard at commit 7a2d171, republished under its MIT licence (© BlackBeltTechnology). 1,058 words, ~2,073 tokens.

Download SKILL.mdSave it as .claude/skills/anti-slop-image-direction/SKILL.md (or your agent's skills folder).
name
anti-slop-image-direction
description
Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for the mockup loop. Images are direction, never source. Never for product UI. Triggers: "generate design references", "image direction for this page", "show me visual directions".
license
MIT
metadata.author
blackbelt-technology
metadata.version
0.1
metadata.adapted_from
Leonxlnx/taste-skill@18dfc92 (image-to-code-skill / image-to-code + imagegen-frontend-web, MIT), inverted: images are direction, not source. Section map…

anti-slop-image-direction

Generated reference images break a model out of its default layouts (centered hero over dark mesh, three equal cards). They are also paid, slow, and full of the same tells we ban. This skill keeps the benefit and drops the risk: images are an opt-in direction aid, distilled into a written brief, and the mockup loop's gates still decide.

Authority: advisory. Findings feed the frontend-mockup-loop FIX step and never override a WCAG-AA or severity gate. The direction brief is one input to GROUND; documented public rules and the WCAG gates still win over anything an image suggests.

When it applies

All of these must hold, otherwise generate nothing:

  1. The declared profile (anti-slop-frontend Design Read) is marketing or new-site.
  2. The user explicitly asked for image direction (or passed an explicit flag).
  3. The user confirmed the plan in step 1.

Image direction is forbidden for product-ui. When asked for a dashboard, admin panel, editor or any app screen, refuse image generation, say that image direction does not apply to product UI, and offer a text-only direction brief instead.

Images are experimental; the text brief is primary

The text-only brief (see Fallback) is the primary, always-available path. Images are an experimental extra. pi-nano-banana ships as a TypeScript bin: a global npm / pi install copy does not run as a plain command (ERR_UNSUPPORTED_NODE_MODULES_TYPE_STRIPPING). Resolve the runner like this:

WhereRunner ($NB)
pi-agent-dashboard monoreponpx --no-install tsx packages/nano-banana/src/bin/nano-banana.ts (the repo's locked tsx)
after pi install npm:@blackbelt-technology/pi-dashboard-nano-banananpx --yes tsx@4.23.12 ~/.pi/agent/npm/node_modules/@blackbelt-technology/pi-dashboard-nano-banana/src/bin/nano-banana.ts (exact pinned tsx)

Probe: run $NB with no arguments; it prints a usage: pi-nano-banana ... line. No usage line → the backend is unavailable → Fallback.

Never run an unpinned npx tsx outside the monorepo: npx would fetch whatever tsx the registry serves and execute it with your permissions.

Backends: gemini (default, needs GEMINI_API_KEY in the environment or a .env) or pi (--backend pi, an OpenRouter credential inside pi; a different bill).

Procedure

1. Plan, then confirm
  • Resolve <mockupDir> the same way the mockup loop's MOCKUP step does: in the pi-agent-dashboard repo, openspec/changes/<name>/mockups/ when a change exists, else mockups/<slug>/; elsewhere, the loop's mockup directory. References go in <mockupDir>/refs/.
  • List the page sections. Plan one image per section.
  • Pick the variation axes for each section (below).
  • Confirm with ask_user, stating: the image count, the backend (gemini or pi), and that generation is paid. No confirmation → generate nothing.
  • Headless rule: interactivity is judged by whether the ask_user tool is in your tool list. If it is absent (subagent, CI, headless run), the run is headless: generate nothing and write the text-only brief.
2. Generate
  • One image per planned section, using only the backend named in the confirmed plan:

    bash
    $NB "<prompt>" --output <mockupDir>/refs/<nn>-<section>.png            # gemini
    $NB "<prompt>" --output <mockupDir>/refs/<nn>-<section>.png --backend pi
  • A backend change requires a new confirmation. If the confirmed backend fails (e.g. Gemini key missing) and the other one is available, stop, re-plan naming the new backend, and ask again. Never switch the backend silently.

  • Re-generating an image, or adding images beyond the confirmed count, also needs a new confirmation.

  • Every prompt carries its section's variation axes and the Part A bans: no purple/violet glow, no neon gradient, no div-style fake dashboard, no stock "Jane Doe" faces, no em-dashes in any rendered text.

  • Describe copy, never quote it: "a short two-line headline about release speed", not the headline itself.

3. Analyze → refs/direction.md

Read every image and write <mockupDir>/refs/direction.md with, per section and for the set:

  • Layout family and composition anchor.
  • Spacing rhythm (tight / regular / airy, section scale).
  • Palette as hex values sampled from the image.
  • Type character (geometric sans, grotesk, mono, editorial serif, ...).
  • Observed tells: every Part A tell seen in an image (e.g. AI-purple glow, fake UI panel, pure black). Observed tells are listed and excluded: they never carry into the palette or layout direction.
  • Consistency across images: one neutral temperature, one accent family.
Show full SKILL.md (430 more words)Show less
4. Hand-off

direction.md becomes one input to frontend-mockup-loop step 1 (GROUND), next to the shipped UI and the documented public rules. Where the direction conflicts with a cited rule or a WCAG gate, the rule and the gate win.

Direction, not source

  • Images are direction, never the source of truth. The mockup is built from the brief plus cited rules, not traced from pixels.
  • Text visible in a generated image is a placeholder and is never transcribed into copy, headings, alt text or the brief.
  • Part A tells are checked on the images themselves, not only on the code.

Variation axes

Pick one value per axis per section; vary across the page. This section is also the text-only brief when no image is generated.

  • Composition anchor: centered statement · top-left lead · bottom-left text over image · left-third caption + two-thirds visual (never twice in a row) · off-grid editorial offset · stacked center · image as canvas with a clean text safe area. At least 3 different anchors per page.
  • Hero scale (per page): giant statement · mid editorial · mini minimalist.
  • Background mode: solid surface with inline asset · subtle texture or grid · full-bleed image with tonal overlay · editorial side image · flat colour block + detail crop · low-chroma tonal gradient · duotone image. Never the same mode on every section.
  • Type character: geometric sans · grotesk · mono accent · editorial serif (only when the brief names one, anti-slop-frontend A2).
  • CTA variation: primary pill · outline · inline link with arrow · full-width banner · caption under a visual. Vary at least once; the primary action stays unmistakable.

Fallback

When the runner is not resolvable, the probe prints no usage line, or the backend has no usable credentials: report the reason in one line, write <mockupDir>/refs/direction.md text-only from the variation axes section above, and continue. Do not fail the task. The same text-only brief is the output of a headless run and of a declined confirmation.

Pitfalls

  • Do NOT generate for product-ui, ever, even when asked twice.
  • Do NOT generate before the user confirmed count, backend and cost.
  • Do NOT copy a headline, label or number out of an image.
  • Do NOT let an image's palette smuggle in a tell (purple glow, pure black).

Verification

  • Profile is marketing or new-site; the request was explicit.
  • The confirmation names image count, backend and paid, and precedes the first image; any backend change has its own confirmation.
  • refs/ holds exactly the confirmed number of images, one per section.
  • refs/direction.md lists layout family, rhythm, palette hex, type character and observed tells; no image text appears verbatim in the mockup.
  • Headless, declined or unavailable runs produced a text-only brief and zero images.

© BlackBeltTechnology, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in packages/anti-slop/.pi/skills/anti-slop-image-direction of BlackBeltTechnology/pi-agent-dashboard.

Open the folder on GitHubat commit 7a2d171

Compare with similar skills

Anti Slop Image Direction next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Anti Slop Image Direction compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Anti Slop Image Direction this skillBlackBeltTechnology/pi-agent-dashboard315—~2.1kAutomated safety check: NotesMIT
WeChat Article Publisherjiji262/wechat-publisher274—~4.4kAutomated safety check: PassNone
Anti Slop Writingadenaufal/anti-slop-writing147—~5.6kAutomated safety check: PassMIT
Anti Slop Writing Idadenaufal/anti-slop-writing147—~9.4kAutomated safety check: PassMIT
Zy Cinematic Realismpopopo-99/zy-cinematic-realism570—~4.6kAutomated safety check: PassCC-BY-NC-4.0
Article Illustration Batch GeneratorSpaceZephyr/design-buddy176—~1kAutomated safety check: PassNone

Similar skills

  • WeChat Article Publisher

    jiji262/wechat-publisher

    Researches a topic, writes an illustrated WeChat Official Account article in a chosen author voice and sends it to the account's draft box.

    274 GitHub stars~4.4k tokensUpdated 2 mo ago
    Writing & ContentAuto-check passed
  • Anti Slop Writing

    adenaufal/anti-slop-writing

    Edit English prose for clarity, specificity, natural flow, and fit to the requested voice.

    147 GitHub stars~5.6k tokensUpdated 14 days ago
    Writing & ContentAuto-check passed
  • Anti Slop Writing Id

    adenaufal/anti-slop-writing

    Tulis dan sunting teks Bahasa Indonesia agar jelas, spesifik, alami untuk audiensnya, dan sesuai suara yang diminta.

    147 GitHub stars~9.4k tokensUpdated 14 days ago
    Writing & ContentAuto-check passed
  • Zy Cinematic Realism

    popopo-99/zy-cinematic-realism

    Develop supplied scripts, short stories, or synopses into scene understanding, director-facing art concepts, and motivated narrative keyframes; compile scene ideas, visual references, or existing…

    570 GitHub stars~4.6k tokensUpdated 7 days ago
    Media & CreativeAuto-check passed
  • Article Illustration Batch Generator

    SpaceZephyr/design-buddy

    Splits a Markdown article by heading, drafts an image prompt for each section, then generates diagrams with the Gemini API and embeds them in the note.

    176 GitHub stars~1k tokensUpdated 3 mo ago
    Media & CreativeAuto-check passed
  • Thumbnail Creator

    mohitagw15856/pm-claude-skills

    Generate article or newsletter thumbnail candidates using the Gemini API from inside Claude Code.

    1.4k GitHub stars~6.3k tokensUpdated 2 days ago
    Media & CreativeAuto-check passed

More from BlackBeltTechnology/pi-agent-dashboard

All 70 skills in this repo
  • Browser

    BlackBeltTechnology/pi-agent-dashboard

    Browser automation via the agent-browser CLI. An agent skill from BlackBeltTechnology/pi-agent-dashboard.

    316 GitHub stars~2k tokensUpdated today
    Auto-check passed
  • CI Troubleshoot

    BlackBeltTechnology/pi-agent-dashboard

    Diagnose failed GitHub Actions runs for pi-agent-dashboard: the 11-file workflow taxonomy, affected-test selection, the release pipeline, known failure modes, and how to read gh run logs and…

    316 GitHub stars~3.5k tokensUpdated today
    Auto-check passed
  • Debug Dashboard

    BlackBeltTechnology/pi-agent-dashboard

    Diagnose problems in the running pi-agent-dashboard system: server.log, /api/health, bridge WebSocket connectivity, vitest triage, known-issue FAQ entries.

    316 GitHub stars~1.6k tokensUpdated today
    Auto-check passed
  • Implement

    BlackBeltTechnology/pi-agent-dashboard

    Disciplined implementation in pi-agent-dashboard: the rebuild matrix (extension→reload, server→restart, client→build+restart, openspec-apply→full rebuild) plus the project's code discipline rules.

    316 GitHub stars~2k tokensUpdated today
    Auto-check passed
  • Pi Dashboard

    BlackBeltTechnology/pi-agent-dashboard

    Monitor and control the pi-dashboard server. An agent skill from BlackBeltTechnology/pi-agent-dashboard.

    316 GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • Session To Guideline

    BlackBeltTechnology/pi-agent-dashboard

    Turn a pi session into a Markdown "how-we-did-it" collaboration guideline: reads the session's JSONL transcript and synthesizes a reusable playbook of which prompts worked, what had to be steered…

    316 GitHub stars~3.2k tokensUpdated today
    Auto-check passed

Questions about Anti Slop Image Direction

What does Anti Slop Image Direction do?

Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for…. Anti Slop Image Direction is an agent skill from BlackBeltTechnology/pi-agent-dashboard. Opt-in, paid reference images for marketing or new-site work: plan one image per section, confirm count, backend and cost, generate with pi-nano-banana, then distill a written direction brief for the mockup loop.

When should I use Anti Slop Image Direction?

Anti Slop Image Direction fits situations like: tasks that involve Humanizing AI text; tasks that involve Image generation.

How do I install Anti Slop Image Direction in Claude Code?

Run `npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-image-direction -a claude-code`. Or copy the skill folder (packages/anti-slop/.pi/skills/anti-slop-image-direction in BlackBeltTechnology/pi-agent-dashboard) into .claude/skills/anti-slop-image-direction in your project. Claude Code loads it when a task matches its description.

How do I install Anti Slop Image Direction in Codex?

Run `npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-image-direction -a codex`. Or copy the skill folder (packages/anti-slop/.pi/skills/anti-slop-image-direction in BlackBeltTechnology/pi-agent-dashboard) into .agents/skills/anti-slop-image-direction in your project. Codex loads it when a task matches its description.

Can I use Anti Slop Image Direction in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-image-direction -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/anti-slop-image-direction, .gemini/skills/anti-slop-image-direction, .github/skills/anti-slop-image-direction and .opencode/skills/anti-slop-image-direction in your project.

What does Anti Slop Image Direction need to run?

Going by SKILL.md and its folder, Anti Slop Image Direction needs the command-line tools its instructions call (npx) and credentials named GEMINI_API_KEY. Our summary lists: Node.js; A credential in GEMINI_API_KEY.

Does Anti Slop Image Direction access the network?

SKILL.md contains no URLs. Its commands use npx, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Anti Slop Image Direction safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Anti Slop Image Direction use?

Anti Slop Image Direction is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Anti Slop Image Direction use?

About 2.1k tokens (SKILL.md is roughly 8.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Anti Slop Image Direction?

Skills that share tags, products or a category with Anti Slop Image Direction: WeChat Article Publisher (jiji262/wechat-publisher, 274 stars), Anti Slop Writing (adenaufal/anti-slop-writing, 147 stars), Anti Slop Writing Id (adenaufal/anti-slop-writing, 147 stars) and Zy Cinematic Realism (popopo-99/zy-cinematic-realism, 570 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Anti Slop Image Direction?

BlackBeltTechnology (a GitHub organization) maintains it in BlackBeltTechnology/pi-agent-dashboard, which has 315 GitHub stars. The repository holds 70 skills in this directory. The repository was last updated on October 10, 2026.

Source: BlackBeltTechnology/pi-agent-dashboard on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.