speak-with-dead
Pass
Audited by Gen Agent Trust Hub on Jun 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely composed of markdown instructions and contains no executable scripts, shell commands, or binary components.
- [NO_CODE]: There are no runtime dependencies or third-party packages required beyond the agent's base capabilities.
- [SAFE]: No evidence of prompt injection, obfuscation, or data exfiltration attempts was found in the instructions or metadata.
- [SAFE]: The skill implements strong logical guardrails by instructing the agent to differentiate between evidence and inference and to explicitly cite sources, which helps mitigate the risks associated with processing untrusted legacy artifacts.
Audit Metadata