The Agent Skills Directory

[PROMPT_INJECTION]: No malicious override patterns or safety bypass instructions were detected. The instructions focus on technical documentation and specification clarity.
[DATA_EXFILTRATION]: No network operations, hardcoded credentials, or access to sensitive file paths were found.
[REMOTE_CODE_EXECUTION]: The skill does not perform any remote code downloads or execution. It consists entirely of markdown and JSON configuration for evaluation.
[OBFUSCATION]: No hidden content, encoded strings, or homoglyph attacks were identified in the text or metadata.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process user-provided architectural requirements, it lacks any high-risk capabilities (e.g., shell access, file writes, network calls) that could be exploited through malicious input. It serves as a text-generation framework for specifications.
[DYNAMIC_CONTEXT_INJECTION]: No use of shell-execution patterns (such as backtick commands) was found in the documentation.

architectural-spec