The Agent Skills Directory

[SAFE]: The skill implements a standard workflow for modifying a configuration file (fields.yaml) used in research outlines. All operations are aligned with the skill's stated purpose of enhancing data collection schemas.
[SAFE]: Data ingestion from external sources, such as user input or the output of a web-searching subagent, is protected by a mandatory human-in-the-loop confirmation step. The agent is explicitly instructed to present all proposed changes to the user for review, editing, or rejection before any file system writes occur.
[SAFE]: The use of the Task tool to invoke a general-purpose subagent for information gathering is a standard use of agent capabilities and does not involve the execution of untrusted external scripts or arbitrary shell commands.
[SAFE]: File access is restricted to locating and updating the fields.yaml file within the project structure, with no attempts to access sensitive system files or credentials detected.
[SAFE]: No indicators of prompt injection, obfuscation, or unauthorized persistence mechanisms were found in the skill's instructions or configuration.

research-add-fields