@1365/stickerforge

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill fetches public NFT images and metadata from IPFS via the listed public gateways (IPFS_GATEWAYS like https://ipfs.io/ipfs in tools._fetch_from_ipfs and the _run_pudgy_pipeline), then merges on-chain metadata into the traits used to build prompts for the image-generation pipeline, so untrusted third‑party content can influence model prompts and generation behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches Pudgy Penguin metadata and images at runtime from the IPFS gateways (https://ipfs.io/ipfs, https://gateway.pinata.cloud/ipfs, https://dweb.link/ipfs), and that fetched on-chain metadata is merged into the trait data which is then injected into the model prompts used for image generation, so the external content directly influences prompts and is a required runtime dependency.