fluentci-ci-cd
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project manifests to automate pipeline execution, creating a surface for indirect instructions. * Ingestion points: Reads configuration from fluentci.toml, .fluentci/workflows/, fluentci-module.toml, and fluentci.lock. * Boundary markers: Instructions advise the agent to inspect jobs before running and establish the actual interface, but do not provide strict programmatic boundaries. * Capability inventory: The skill can execute shell commands via fluentci run and Wasm modules with host access (filesystem, network, deployments). * Sanitization: No explicit content sanitization or validation is defined for manifest data.
- [REMOTE_CODE_EXECUTION]: The skill documents the use of a shell-piped installation script (curl | sh) from the vendor's official domain.
- [REMOTE_CODE_EXECUTION]: The fluentci run command allows for the download and execution of Wasm modules from remote HTTP(S) URLs, which can perform operations on the host system.
- [EXTERNAL_DOWNLOADS]: Downloads CLI components, Wasm modules, and dependencies from official vendor infrastructure and repositories.
- [COMMAND_EXECUTION]: Executes various CLI commands such as fluentci run, fluentci workflows, and fluentci modules to manage and run CI/CD tasks.
Audit Metadata