olares-publish
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, potentially untrusted data which could serve as a vector for indirect prompt injection or data poisoning.
- Ingestion points: The agent is instructed to fetch app logos from external repositories and websites, capture screenshots from running instances, and read or translate app manifest files (OlaresManifest.yaml). It also monitors GitBot labels and PR status on GitHub.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external manifest content or GitBot messages as untrusted data.
- Capability inventory: The skill possesses the ability to write files to the local system (processed icons and images), perform network requests to the OpenAI API (for promotional image rendering), and interact with the GitHub API for PR management.
- Sanitization: While the Python scripts perform basic image validation (size, format), the textual content of the manifests is processed by the AI without specific sanitization filters.
- [EXTERNAL_DOWNLOADS]: The documentation references several external dependencies required for its workflows. These include the @beclab/olaresid Node.js package for identity management, and Python packages such as Pillow, cairosvg, numpy, and realesrgan for image processing. These packages are standard for the skill's stated functionality.
Audit Metadata