red-team-tactics
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: LOW
Full Analysis
- PROMPT_INJECTION (SAFE): No instructions found that attempt to override agent behavior or bypass safety constraints. The content is descriptive and educational in nature.
- DATA_EXFILTRATION (SAFE): No hardcoded credentials, sensitive file paths, or network operations (curl, wget) were detected.
- REMOTE_CODE_EXECUTION (SAFE): The skill does not perform any external downloads or execute remote scripts.
- COMMAND_EXECUTION (SAFE): No shell commands or system-level execution patterns are present. The 'allowed-tools' (Read, Glob, Grep) are standard for file inspection and are used here for a reference guide.
- OBFUSCATION (SAFE): No encoded strings, homoglyphs, or hidden characters were found in the text.
- INDIRECT_PROMPT_INJECTION (SAFE): While the skill uses tools to read files, its primary function is informational and it does not ingest untrusted data from external sources for processing or decision-making.
Audit Metadata