The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions do not contain any patterns attempting to bypass safety filters or override the AI's core instructions. It explicitly enforces ethical constraints, such as 'Hard Rules' against fabricating statistical results or ethics approval IDs.
[DATA_EXFILTRATION]: No commands for accessing sensitive user files (e.g., SSH keys, credentials) or performing network exfiltration were found. The skill operates purely within the text-processing context provided by the user.
[CREDENTIALS_UNSAFE]: There are no hardcoded API keys, secrets, or authentication tokens within the skill files.
[REMOTE_CODE_EXECUTION]: The skill does not perform any dynamic code execution or download external scripts. It relies on static markdown reference files for its knowledge base.
[EXTERNAL_DOWNLOADS]: References to external URLs are limited to official and well-known scientific resources, such as the EQUATOR Network and specific reporting guideline organizations (e.g., consort-statement.org). These are informational references and do not involve executable content.
[SAFE]: The skill is well-documented and restricts itself to academic writing assistance. It employs a placeholder system ([AUTHOR TO SPECIFY: ...]) to handle missing information safely without fabricating data.

methods-section-writer