The Agent Skills Directory

[SAFE]: The skill is a legitimate writing tool that follows secure repository management practices. No malicious code, exfiltration logic, or privilege escalation patterns were identified.
[DATA_EXFILTRATION]: The skill reads existing blog files from standard content directories (e.g., src/content/blog/) to extract stylistic 'voice tokens'. This access is strictly bounded to the skill's stated purpose of matching the author's voice and does not target sensitive system or credential files.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes data from user briefs and external research agents. This is mitigated by the skill's intensive editorial workflows.
Ingestion points: Untrusted data enters the agent context through user-provided briefs and reports from research sub-agents (as detailed in references/modes.md).
Boundary markers: No explicit boundary markers or 'ignore' instructions for the research data are defined in the prompt templates.
Capability inventory: The skill possesses file-read and file-write capabilities restricted to project-specific content and planning directories (detailed in references/seo-and-files.md).
Sanitization: The skill mandates a rigorous three-stage humanization and 'soul check' process (Stage 6) that audits the draft for AI tells and verifies claims, effectively acting as a quality-focused sanitization layer for ingested research.

editorial-engineer