writing-hookify-rules
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional and does not include any executable code, remote dependencies, or obfuscation. It focuses on providing templates and guidance for defensive monitoring.
- [PROMPT_INJECTION]: The skill documents a surface for indirect prompt injection by describing how the agent ingests and follows instructions from project-local files. However, the documentation specifically targets the creation of safety guardrails, and this surface is inherent to the intended defensive purpose of the tool.
- Ingestion points: Rules are stored in project-local markdown files within the
.claude/directory. - Boundary markers: The system uses YAML frontmatter delimiters to separate configuration from the message body.
- Capability inventory: The documented rules can monitor and potentially block bash commands, file edits, and agent stop events.
- Sanitization: The skill provides guidance on writing regex patterns but does not include automated sanitization logic, as the files are intended for project-specific customization.
Audit Metadata