The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: No external downloads or remote script executions were identified. The skill relies entirely on local reference files within its own directory.
[REMOTE_CODE_EXECUTION]: The skill does not contain any code execution patterns, package installations, or dynamic script generation.
[DATA_EXFILTRATION]: There are no network operations or commands that attempt to access and exfiltrate sensitive user data or system credentials.
[PROMPT_INJECTION]: The skill uses detailed persona instructions (e.g., acting as a cognitive psychology expert) but does not include any attempts to bypass model safety filters or override core system instructions.
[COMMAND_EXECUTION]: No shell commands or system-level operations were detected. The skill's operations are limited to text processing and referencing local markdown files.
[OBFUSCATION]: The content is presented in clear text. No hidden characters, encoded URLs, or homoglyph-based attacks were found.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests user-provided text (e.g., a draft post or a topic description) to perform analysis and title generation.
Boundary markers: No explicit delimiters or boundary instructions are used to isolate user input from the rest of the prompt logic.
Capability inventory: The skill possesses no dangerous tools or capabilities; it only outputs structured markdown text.
Sanitization: No input sanitization or validation is present, but given the lack of executable capabilities, the risk is negligible.

li-cover