sglang-bisect-ci-regression
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of the GitHub CLI (
gh),git,ssh, anddockerto interact with CI environments, repository history, and remote hosts for analysis and reproduction. - [EXTERNAL_DOWNLOADS]: Fetches source code directly from the official SGLang GitHub repository and installs dependencies from standard package registries during environment setup.
- [REMOTE_CODE_EXECUTION]: Generates minimal Python reproduction scripts based on analyzed failure signatures and executes them on user-provided remote targets via SSH and Docker.
- [PROMPT_INJECTION]: Ingests external data from remote CI logs and repository metadata. This creates an indirect prompt injection surface as the agent processes potentially untrusted log content to generate reproduction scripts without explicit sanitization or boundary markers.
Audit Metadata