product-photo-studio — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited product-photo-studio (Agent Skill) and scored it 87/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 3 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 3 flagged
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 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.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
A single skill that covers every "transform a product photo with AI" workflow Picsart's gen-ai CLI supports. Use this whenever the user has product photography (single packshot, full catalog, or one artwork) and needs it re-rendered, re-staged, or fanned out into variants. Replaces six narrower skills with one entry point and six mode references.
Input: one or more product photos. Output: styled, composed, or fanned-out variants ready for PDPs, marketplaces, ads, mockup listings, or campaigns.
Pick the mode that matches the task. If the user's request maps to any of these, this skill is the right one:
| Mode | Trigger phrases | Reference |
|---|---|---|
| bulk-restyle | "catalog styling", "consistent background across SKUs", "PDP-ready staging at scale" | references/modes/bulk-restyle.md |
| compose | "lifestyle compose", "product in context", "put this product in a scene" | references/modes/compose.md |
| seasonal | "seasonal refresh", "holiday catalog", "Christmas / summer / Black Friday catalog" | references/modes/seasonal.md |
| variants | "color variants", "material variants", "colorway fan-out", "size chart re-render" | references/modes/variants.md |
| reshoot | "batch reshoot", "regenerate catalog", "enterprise catalog re-render with brand rules" | references/modes/reshoot.md |
| mockups | "product mockup", "POD mockup", "Etsy / Shopify listing mockup", "in-hand render" | references/modes/mockups.md |
If the user's task involves video, characters, or persona generation, this is the wrong skill — see motion-studio, gen-ai-persona-creation, gen-ai-use.
Picsart gen-ai CLI installed and authenticated:
# Install (signed binary, recommended)
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
# Authenticate
gen-ai login
gen-ai whoami # verifyPer-mode prerequisites (image counts, brand files, manifests) are documented inside each mode reference. Always confirm pricing before a bulk run:
gen-ai pricing --model <model> --count <N>Read references/modes/<mode>.md.# Single image (compose / mockups)
gen-ai generate --model <model> --image input.jpg --prompt "<prompt>"
# Batch (bulk-restyle / seasonal / variants / reshoot)
gen-ai batch --manifest manifest.json
# Estimate cost before running
gen-ai pricing --model <model> --count <N>
# Browse available models
gen-ai modelsManifest patterns, model recommendations, and per-mode best practices live in the individual mode references.
Always follow the same outer loop regardless of mode:
gen-ai pricing before committing. Surface the total to the user.gen-ai generate or gen-ai batch. Stream progress.Mode-specific pitfalls (e.g. "shadow direction drifts across variants", "seasonal overlays expire", "brand.md gate blocks SKU-XYZ") live inside each mode reference. Shared pitfalls:
enterprise-brand-governor-style brand.md in scope, the manifest must reference it.enterprise-pinned-registry.After any run:
# Confirm output count matches expected
ls -1 outputs/<run>/ | wc -l
# Spot-check one output's dimensions
gen-ai inspect outputs/<run>/<sample>.jpgIf anything looks off, re-run with --debug and consult the mode reference's "Common pitfalls" section.
motion-studio — video pipeline (out of scope here)gen-ai-use — foundational gen-ai CLI referenceenterprise-brand-governor — policy gatingenterprise-pinned-registry — version pinning~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.