imagevideogen 87/100 — SaferSkills

CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

22.4 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 3 flagged

Securityscore 64 · 3 findings

MEDIUMInstruction telling the agent not to ask for approvalSS-SKILL-INJECT-DONT-ASK-01 · Prompt injection · prompts/replicate/prompting-guide.md×2

MEDIUM — it fires on intent; the real damage depends on the host agent's own approval-gating.

Why it matters

The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.

The exact value spotted

excerptprompts/replicate/prompting-guide.md· markdown

78### UI Anti-patterns

80- Do not prompt only with style words and no product screenshot reference

81- Do not ask for a vague `SaaS dashboard` if you want the real app preserved

82- Do not omit the constraint `Do not redesign the UI`

Occurrences

2 occurrences · first at L80, also L81

Show all 2 locations

Line

File

L80

prompts/replicate/prompting-guide.md

L81

prompts/replicate/prompting-guide.md

How to fix

Remove the approval-skipping instruction, or scope it narrowly to a specific safe, reversible action.

Delete blanket "don't ask / no need to confirm" directives from the skill.
If the skill is a genuine autonomous job, restrict the opt-out to a named non-destructive action rather than all actions.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-DONT-ASK-01sha2568169c469ab5917e7…rubric 365aacaView on GitHub

MEDIUMInstruction telling the agent not to ask for approvalSS-SKILL-INJECT-DONT-ASK-01 · Prompt injection · prompts/replicate/ui-design-guide.md

MEDIUM — it fires on intent; the real damage depends on the host agent's own approval-gating.

Why it matters

The exact value spotted

excerptprompts/replicate/ui-design-guide.md· markdown

18- keep prompt text concrete and short

19- let reference images carry style and composition

20- never ask one prompt to invent both the brand system and the presentation system

22## What to borrow from the article

Occurrences

1 occurrence · at L20

How to fix

Remove the approval-skipping instruction, or scope it narrowly to a specific safe, reversible action.

Delete blanket "don't ask / no need to confirm" directives from the skill.
If the skill is a genuine autonomous job, restrict the opt-out to a named non-destructive action rather than all actions.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-DONT-ASK-01sha25614b7852251bf4666…rubric 365aacaView on GitHub

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

prompts/replicate/skill.md · 1 file

prompts/replicate/skill.md5.8 KB · Markdown

name: imagevideogen

description: >

Generate images and videos using multiple AI models via Replicate.

Supports text-to-image (flux-pro, flux-schnell, sdxl, recraft-v3),

text-to-video and image-to-video (kling, minimax, luma, veo),

with built-in prompting best practices and presets.

Trigger: generate image, generate video, create image, create video,

text to image, text to video, image to video, t2i, t2v, i2v,

imagevideogen, image gen, video gen, make an image, make a video,

logo, icon, logo design, brand mark, logo generation, logo inspiration

tags: [image, video, generation, replicate, ai, creative, flux, kling, veo, logo, icon, brand]

mcp_servers: []

<overview> Multi-model image and video generation skill using Replicate API.

Image models: flux-pro (photorealism), flux-schnell (fast/cheap), sdxl (versatile), recraft-v3 (vector/logos) Video models: kling (portraits), minimax (atmospheric), luma (physics-aware), veo (cinematic+audio)

All generation runs through a single generate.py script with zero external dependencies. Prompting presets and best practices are built in.

For UI/product visuals, use the screenshot-derived prompt rules in:

prompts/replicate/ui-design-guide.md

Key rule:

the real product screenshot is the hero asset
style references only control framing and polish
do not redesign the UI unless explicitly requested

</overview>

<check_setup> Verify REPLICATE_API_TOKEN is available:

source ~/.claude/.env 2>/dev/null
if [ -z "$REPLICATE_API_TOKEN" ]; then
  echo "MISSING: Add REPLICATE_API_TOKEN to ~/.claude/.env"
else
  echo "OK: Token found (${REPLICATE_API_TOKEN:0:8}...)"
fi

If missing, get a token at https://replicate.com/account/api-tokens and add to ~/.claude/.env:

REPLICATE_API_TOKEN=r8_your_token_here

</check_setup>

<intake> What would you like to create?

Generate an image — Text-to-image with model selection
Generate a video — Text-to-video from a description
Animate an image — Image-to-video (provide a source image)
Design a logo/icon — Logo generation with recraft-v3 + inspiration browsing
Quick draft image — Fast cheap image with flux-schnell ($0.003)
List models & pricing — See all available models
Prompting help — Best practices for quality results

</intake>

Choice	Workflow	Default Model
1. Generate image	workflows/image.md	flux-pro
2. Generate video	workflows/video.md	kling
3. Animate image	workflows/video.md (with --image)	kling
4. Logo/icon design	workflows/logo.md	recraft-v3
5. Quick draft	workflows/image.md	flux-schnell
6. List models	Run: `python3 generate.py image -l` and `python3 generate.py video -l`	—
7. Prompting help	references/prompting-guide.md	—

Auto-detection: If the user's request mentions logo, icon, brand mark, or logotype, automatically route to workflows/logo.md — do not use flux-pro for logo work. </routing>

<quick_reference>

# Source credentials
source ~/.claude/.env

# Image generation
python3 .claude/skills/imagevideogen/scripts/generate.py image \
  -m flux-pro -p "Mountain landscape, golden hour, dramatic clouds" -o mountain.webp

# Fast draft ($0.003)
python3 .claude/skills/imagevideogen/scripts/generate.py image \
  -m flux-schnell -p "Logo design for a tech startup" --preset logo -o logo.webp

# Video from text
python3 .claude/skills/imagevideogen/scripts/generate.py video \
  -m kling -p "Gentle ocean waves on a sandy beach, slow motion" -d 5 -o waves.mp4

# Animate an image
python3 .claude/skills/imagevideogen/scripts/generate.py video \
  -m kling -p "Subtle head nod and gentle smile" -i portrait.png -o animated.mp4

# Logo/icon design (always use recraft-v3)
python3 .claude/skills/imagevideogen/scripts/generate.py image \
  -m recraft-v3 -p "Minimalist geometric logo mark of a column, flat colors" --preset logo -o logo.png

# List available models
python3 .claude/skills/imagevideogen/scripts/generate.py image --list-models
python3 .claude/skills/imagevideogen/scripts/generate.py video --list-models

</quick_reference>

<model_comparison>

Image Models

Model	Cost	Speed	Best For
flux-pro	$0.04/img	~5s	Photorealism, general purpose, highest quality
flux-schnell	$0.003/img	~1s	Fast drafts, iteration, low cost
sdxl	$0.008/img	~8s	Versatile, wide style range, LoRA support
recraft-v3	$0.04/img	~5s	Vector art, logos, icons, illustrations

Video Models

Model	Cost	Duration	Best For
kling	$0.07/s	5-10s	Silent video, micro-expressions, portraits
minimax	$0.12/s	5s	Atmospheric scenes, cinematic quality
luma	$0.032/s	5s	Physics-aware motion, nature, fast gen
veo	$0.40/s	5-8s	Highest quality, dialogue audio generation

</model_comparison>

<reference_index>

references/prompting-guide.md — Image and video prompting best practices
references/ui-design-guide.md — UI/product visual prompt recipe adapted from the UX Planet article
references/models.md — Detailed model comparison and parameters
references/troubleshooting.md — Common errors and fixes

</reference_index>

<workflows_index>

workflows/setup.md — First-time setup and token configuration
workflows/image.md — Image generation walkthrough
workflows/video.md — Video generation and animation walkthrough
workflows/logo.md — Logo/icon design with recraft-v3 + logosystem.co inspiration

</workflows_index>

<success_criteria>

Output file exists and is non-empty
Image dimensions match requested (or model default)
Video plays without corruption
Cost logged for user awareness
No API errors in stderr

</success_criteria>

imagevideogen— agent skill

Is imagevideogen safe to install?

Image Models

Video Models

Audit the pieces. Scan the whole. Decide.