feature-release-rollout
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches source code from the official OpenHands GitHub organization for analysis. Evidence: git clone commands for software-agent-sdk, OpenHands-CLI, OpenHands, docs, and growth-utils.
- [INDIRECT_PROMPT_INJECTION]: The skill reads external codebase content to generate implementation proposals and GitHub issues, representing a standard ingestion surface. Ingestion points: SDK code, documentation, and blog post files from the cloned OpenHands repositories. Boundary markers: Absent; there are no explicit delimiters or warnings to ignore instructions inside the analyzed files. Capability inventory: Command execution (git, grep, gh, npm, make) and GitHub issue creation. Sanitization: Absent; content from the repositories is processed by the agent to create summaries and issue bodies without explicit escaping.
Audit Metadata