The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions found that attempt to bypass AI safety filters or override system behavior.
[Data Exposure & Exfiltration] (SAFE): No sensitive file access or network operations detected. The code only performs local data generation and file writes for test fixtures.
[Obfuscation] (SAFE): No encoded strings, hidden characters, or homoglyphs identified.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): Uses standard, well-known libraries (pandas, PyYAML). It correctly uses yaml.safe_load() to prevent YAML deserialization attacks. No remote code execution patterns or dynamic code execution (eval/exec) were found.
[Persistence & Privilege Escalation] (SAFE): No attempts to modify system configuration, acquire root privileges, or establish persistence across sessions.
[Indirect Prompt Injection] (SAFE): While the skill ingests a YAML schema, it uses yaml.safe_load and maps inputs to discrete data generation functions. There is no path for untrusted input to influence agent instructions or execute commands.

test-fixture-generator