embedded-agency

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No evidence of prompt injection or instructions to bypass safety guidelines was found. The skill maintains a purely educational tone.
  • [DATA_EXFILTRATION]: No sensitive file paths, hardcoded credentials, or network exfiltration patterns were detected. There are no tools or scripts present to perform such actions.
  • [REMOTE_CODE_EXECUTION]: The skill does not contain any executable scripts, package manager commands, or remote download patterns. It consists entirely of markdown and JSON data.
  • [OBFUSCATION]: No obfuscated content, such as base64-encoded strings, zero-width characters, or homoglyphs, was detected in the provided files.
  • [COMMAND_EXECUTION]: There are no shell commands, dynamic context injections, or subprocess calls present in the skill files.
  • [CREDENTIALS_UNSAFE]: No API keys, tokens, or private secrets were found within the skill's metadata or instructional content.
  • [INDIRECT_PROMPT_INJECTION]: The skill does not possess an attack surface for indirect injection as it does not ingest external untrusted data or provide tools to process such data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 08:36 PM
Security Audit — agent-trust-hub — embedded-agency