writing-audit

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill follows a highly structured and modular design, separating routing logic (references/intent-router.csv), playbooks (references/playbooks/), and output templates (templates/). This architecture promotes predictable behavior and limits the risk of prompt-induced errors.
  • [COMMAND_EXECUTION]: A shell script for development-time static analysis is included in the package. (File: evals/run-static-checks.sh). This script executes local commands such as python3 "$repo_root/scripts/check-skill-static.py" to verify skill configuration. These commands are intended for internal developer testing and are not triggered during standard user interaction.
  • [SAFE]: The skill processes user-provided drafts for auditing. This data ingestion surface is well-contained through the use of specific intent playbooks and strict output templates, minimizing the risk of indirect prompt injection affecting agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 05:45 PM
Security Audit — agent-trust-hub — writing-audit