building-edgespark-apps
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill content consists of legitimate documentation and workflow instructions for developing applications using the EdgeSpark CLI and SDKs.
- [PROMPT_INJECTION]: Static analysis identified potential concealment patterns; however, these are evaluated as security guardrails. The instructions explicitly direct the agent to ensure that user secrets (e.g., API keys) never enter the agent context or conversation logs, which is a defensive best practice.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it is designed to ingest and process project files like
edgespark.tomland markdown-based agent instruction files. This risk is inherent to development-assistant skills and is mitigated by the skill's design, which mandates human browser action for sensitive operations like secret management. - [CREDENTIALS_UNSAFE]: The skill correctly implements secret management by using a platform-specific CLI command (
edgespark secret set) and instructing the agent to hand off the final value entry to a human through a secure URL, rather than handling raw secrets itself.
Audit Metadata