muapi-photo-pack-generator — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited muapi-photo-pack-generator (Agent Skill) and scored it 45/100 (orange). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 1 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.
*.sig, SIGNATURES) outside the documentation.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.
Transform a single reference photo into a collection of themed images while maintaining extremely high facial identity fidelity.
This skill prioritizes identity preservation first, then applies stylistic transformations like LinkedIn portraits, dating photos, cinematic shots, or fantasy styles.
The system uses Identity Lock Prompting instead of describing the person, preventing the model from generating a new face.
The generated images must always depict the same person from the reference image.
All prompts MUST include identity lock instructions.
Required identity rules:
The agent MUST analyze the reference image before generation.
However the analysis must NOT describe the person (age, ethnicity, hair etc).
Allowed analysis fields:
Example:
Head orientation: slight left tilt Expression: neutral friendly Lighting: soft frontal light Framing: head and shoulders portrait
Ensure the user has provided a reference image.
Supported inputs:
Extract scene attributes only.
DO NOT describe:
Identity must come directly from the image.
If the user does not specify a category suggest:
Use the reference image as the identity source.
Preserve the exact facial identity from the reference image.
Identity must remain identical to the reference photo.
Do not change:
Maintain similar head orientation as the reference.
Scene example:
Outdoor café portrait Soft natural daylight 35mm portrait lens Shallow depth of field Photorealistic skin texture
Always include:
different person altered face changed facial features new identity generic face beautified face plastic skin face distortion
Example:
bash scripts/generate-pack.sh \ --image "./my_face.jpg" \ --category "LinkedIn" \ --identity-lock true \ --num 5
| Category | Best For | Aesthetic |
|---|---|---|
| Professional | Studio | |
| CEO | Founders | Office |
| Tinder | Dating | Lifestyle |
| OldMoney | Luxury | Estate |
| Cyberpunk | Fantasy | Neon |
| Fitness | Gym | Athletic |
| Travel | Social | Bali/Paris |
| 90s | Retro | Vintage |
| Holiday | Seasonal | Festive |
Identity preservation is always more important than style.
Avoid prompts like:
"Indian man in his 20s with short hair"
This causes the model to generate a new face.
Identity must come from the reference image only.
Best results with:
This system produces:
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.