writing-skills

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The instructions and the 'Persuasion Principles' document describe the use of 'Authority' and 'Commitment' principles to enforce adherence to the TDD methodology. While the skill uses forceful language (e.g., 'YOU MUST', 'No exceptions', 'Delete means delete'), these directives are aimed at maintaining engineering discipline and preventing agents from skipping quality-control steps, rather than bypassing safety filters or content guidelines.
  • [EXTERNAL_DOWNLOADS]: The 'anthropic-best-practices.md' file contains informational links to visual documentation assets hosted on 'mintcdn.com', a well-known content delivery network for documentation platforms. These references are used solely for visual aid and do not involve executable code or untrusted script downloads.
  • [COMMAND_EXECUTION]: The documentation references standard development tools such as git, the GitHub CLI (gh), and common utilities like wc and grep. These are used in the context of typical software development and documentation workflows (e.g., verifying word counts, checking PR status, or committing skill files).
  • [SAFE]: The skill represents a legitimate effort to apply software engineering best practices to agent-driven documentation workflows. It focuses on verification, loop closure, and systematic testing to improve the reliability of AI agent performance.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 09:03 AM
Security Audit — agent-trust-hub — writing-skills