The Agent Skills Directory

[DATA_EXPOSURE_EXFILTRATION]: No network requests, file system access, or credential harvesting patterns were detected. The skill operates entirely within the text processing context.
[REMOTE_CODE_EXECUTION]: The 1 file analyzed is pure Markdown and contains no instructions to install or execute external packages or scripts.
[PROMPT_INJECTION]: The skill uses detailed instructional logic to guide the model's creative output, but contains no patterns designed to bypass safety filters or override system-level constraints.
[OBFUSCATION]: Analysis confirmed the absence of Base64, zero-width characters, homoglyphs, or other techniques used to hide malicious intent.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user-supplied content (drafts, themes, and feature lists). While this represents a data ingestion surface, the skill lacks any capabilities (such as shell execution, network requests, or file writes) that would allow an injection to result in privilege escalation or data exfiltration.
[COMMAND_EXECUTION]: No shell commands or system-level utility calls are present in the skill instructions.

post-optimizer