oh-my-openagent-orchestration
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the 'oh-my-opencode' package globally using 'npm install -g' or execute it directly via 'npx'. Additionally, it fetches documentation from external GitHub repositories maintained by 'code-yeongyu'.
- [COMMAND_EXECUTION]: The skill executes shell commands via 'npx' and provides the agent with capabilities to modify the local filesystem through 'edit_file' and perform workspace-wide operations using LSP-integrated tools like 'lsp_rename' and 'ast_grep'.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by fetching web search results (Exa), documentation (Context7), and code snippets (Grep.app). This content is fed into an autonomous orchestration loop ('ultrawork') that has the capability to write to the codebase, creating a vulnerability surface where external content could influence agent behavior. Mandatory Evidence Chain: (1) Ingestion points: External content fetched via Exa, Context7, and Grep.app. (2) Boundary markers: Not explicitly defined in instructions to separate external data from system instructions. (3) Capability inventory: Full file writing ('edit_file'), LSP workspace modifications, and tool execution ('team_spawn', 'background_agent'). (4) Sanitization: No explicit sanitization of external content is mentioned.
Audit Metadata