The Agent Skills Directory

[SAFE]: The skill is composed of Markdown documentation and reference guides. It does not include any scripts, binary files, or hidden executable content.
[DATA_EXFILTRATION]: No network access or data exfiltration patterns were detected. The skill's operations are confined to the local project directory, specifically interacting with draft manuscripts and an audit tracker.
[COMMAND_EXECUTION]: The skill uses the standard str_replace tool for surgical text modification. No attempts to execute arbitrary shell commands, escalate privileges, or access sensitive system files were found.
[PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface as it ingests untrusted manuscript drafts; however, the risk is negligible due to the diagnostic nature of the work and the requirement for explicit human review before any changes are applied. No markers of malicious intent or safety-bypass instructions are present.

prose-mechanics