astrale-services
Pass
Audited by Gen Agent Trust Hub on Jul 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute repository-specific static checks and tests as part of the verification workflow for new or modified code.
- [REMOTE_CODE_EXECUTION]: The core functionality involves deploying user-provided source code as a Cloudflare Worker and subsequently invoking its functions for end-to-end verification.
- [PROMPT_INJECTION]: There is an inherent surface for indirect prompt injection, as the skill processes user-supplied code and business requirements which could contain instructions designed to manipulate the agent's behavior during the deployment or verification process.
- Ingestion points: Processes existing codebases, workspace paths, and business requirements files (
deployer-agent.md). - Boundary markers: The skill uses instruction headers but does not explicitly define markers for untrusted content.
- Capability inventory: Capable of running subprocesses (tests/checks), network deployment, and function invocation.
- Sanitization: The skill contains instructions to explicitly avoid embedding secrets in payloads or logs, but does not specify input sanitization for the code logic itself.
Audit Metadata