The Agent Skills Directory

[SAFE]: The skill's primary function is to read from and write to project-specific documentation files (INTENT.md and intent/*.md) based on conversation history. This behavior is consistent with its stated purpose of managing project intent.
[SAFE]: No patterns associated with data exfiltration, remote code execution, or credential theft were identified. The skill does not perform network operations or access sensitive system directories.
[SAFE]: The skill utilizes the disable-model-invocation: true configuration, which acts as a security measure by preventing the agent from autonomously calling other LLM models during its execution.
[SAFE]: Although the skill processes untrusted data from the conversation history (Category 8 surface), the risk is mitigated because the resulting output is stored as static markdown documentation without being passed to executable sinks or network tools.

to-intent