The Agent Skills Directory

[PROMPT_INJECTION]: No patterns intended to override agent instructions or bypass safety guidelines were found. The instructions focus on scientific measurement discipline.\n- [CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or sensitive credentials were detected in the skill's code or documentation.\n- [EXTERNAL_DOWNLOADS]: The skill does not reference or download external scripts or packages from untrusted sources. It mentions well-known technical documentation sources for methodology.\n- [OBFUSCATION]: Analysis did not reveal any hidden text, encoded commands, or homoglyph attacks. The content is clear and instructional.\n- [COMMAND_EXECUTION]: The verification steps include standard GCC compilation and execution commands for the skill's own internal C examples. No dangerous command patterns or shell injections were found.\n- [INDIRECT_PROMPT_INJECTION]: While the skill analyzes user-provided C code and performance claims, it defines a rigid verification framework that requires empirical evidence, which naturally mitigates the risk of processing malicious instructions in data.\n- [DYNAMIC_CONTEXT_INJECTION]: The skill manifest does not utilize dynamic context injection patterns to execute shell commands at load time.

performance-measurement-discipline