ip-as-logo
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
- [METADATA_POISONING]: A discrepancy exists between the reported author (bingwu2003) and the identity found in the repository's metadata (s1dashu). The README.md, LICENSE, and CLI installation instructions all attribute the skill to s1dashu, which does not match the expected vendor naming patterns and could mislead users about the skill's origin.
- [INDIRECT_PROMPT_INJECTION]: The skill's workflow (SKILL.md, Step 2) requires the agent to read and process content from workspace files including READMEs, product documentation, and package manifests to extract product context. This ingestion of untrusted data lacks explicit boundary markers or sanitization, creating a surface where a malicious file in the repository could inject instructions to influence the agent's behavior or its image-generation parameters.
- [CREDENTIALS_UNSAFE]: The instructions (SKILL.md, Step 8) direct the agent to ask the user to provide an API key in the chat interface if a suitable image model is not available. Encouraging users to input sensitive credentials directly into a chat window is a poor security practice that can lead to credential exposure in chat history and telemetry logs.
- [EXTERNAL_DOWNLOADS]: The README.md instructs the user to install the skill using the command
npx skills@latest add s1dashu/ip-as-logo-skill. This involves downloading and executing code from an external GitHub repository (s1dashu) which is not identified as a trusted source, presenting a potential supply chain risk.
Audit Metadata