The Agent Skills Directory

[DATA_EXPOSURE]: The skill instructs the agent to read and write to project-specific files such as context/output-config.md, .drafting-output.md, and session logs. These operations are limited to the local project directory for state persistence and do not target sensitive system files or credentials.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided input (descriptions of writing blocks). While it lacks explicit sanitization instructions for this input, it possesses no high-risk capabilities—such as network access or command execution—that could be exploited via an injection attack. The ingestion surface is limited to dialogue and metadata related to writing progress.
[PERSISTENCE]: The skill implements logic to persist session progress and diagnostic notes in local files. This is a functional feature for maintaining context across agent sessions and does not involve malicious persistence mechanisms like modifying shell configurations or system services.
[SAFE]: No evidence of prompt injection, obfuscation, remote code execution, or unauthorized privilege escalation was detected. The skill operates entirely within the scope of a writing diagnostic assistant.

drafting