niblet
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were identified in the skill's instructions or configuration.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill implements secure secret management practices for its API key (
NIBLET_TOKEN). It explicitly forbids the agent from requesting keys via chat and recommends using secure environment variables or host-managed secret stores to prevent credential leakage. - [INDIRECT_PROMPT_INJECTION]: The skill includes robust safeguards against indirect prompt injection. It specifically instructs the agent to treat all data retrieved from external sources (such as design references and catalogue metadata) as untrusted evidence, never as executable instructions or scope-altering commands. This mitigation is explicitly documented in the Reference Evidence Policy.
- [EXTERNAL_DOWNLOADS]: The skill utilizes official vendor resources, including the
@pymodel/nibletNode.js package and theniblet-api.pymodel.comAPI. These originate from the vendor's own infrastructure and are documented as part of the intended product functionality. - [DYNAMIC_EXECUTION]: The
get_ui_componenttool retrieves component source code from a remote registry. The instructions provide clear guidance for safe integration, emphasizing manual adaptation to the local codebase and adherence to existing project tokens rather than arbitrary execution.
Audit Metadata