zig-cli
Fail
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill implements a high-risk installation workflow that downloads a script from a remote URL and executes it immediately using a shell pipe. This pattern bypasses local security inspections and could lead to arbitrary code execution if the remote source is compromised.
- Evidence:
curl -fsSL https://christianhelle.com/openapi2zig/install | INSTALL_DIR="$HOME/.local/bin" bashfound inSKILL.md. - [EXTERNAL_DOWNLOADS]: The skill fetches and executes content from an external personal domain that is not a recognized trusted service or organization.
- Evidence:
https://christianhelle.com/openapi2zig/installreferenced in both the bootstrap and update workflows. - [METADATA_POISONING]: The documentation contains references to absolute file system paths from the developer's local environment, which may leak internal information about the development environment.
- Evidence:
/Users/tbrewer/projects/goodway/henry/cli-zig/appearing inreferences/templates.md. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and processing external OpenAPI specifications which could contain malicious instructions designed to influence the agent's code generation.
- Ingestion points: The skill reads OpenAPI JSON or YAML files from paths specified by the user or common project locations.
- Boundary markers: None; there are no instructions to the agent to ignore or delimit instructions found within the spec files.
- Capability inventory: The skill has extensive capabilities including file system modification, command execution (Zig compiler, shell), and network operations.
- Sanitization: None; the agent is instructed to perform complex manual fixes and transformations on the external data without validation of the input text.
- [COMMAND_EXECUTION]: The skill performs multiple shell operations including project initialization, build processes, and cross-compilation workflows.
- Evidence: Commands such as
zig init,zig build,rm,cp, andshasumare used acrossSKILL.md.
Recommendations
- AI detected serious security threats
Audit Metadata