meta-prompting-suite

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The suite uses the Bash tool to perform administrative operations such as directory listing and file copying within the agent's configuration environment (~/.claude/).- [EXTERNAL_DOWNLOADS]: Installation instructions include cloning the skill repository from the author's GitHub account (HermeticOrmus), which is a vendor-owned resource.- [REMOTE_CODE_EXECUTION]: The skill can generate and write new slash commands and agent definitions. While these can contain executable logic, the skill includes manual verification steps and a secret scanning mechanism to mitigate risk.- [PROMPT_INJECTION]: The skill analyzes session history and external agent files to identify functional gaps or redundancies.
  • Ingestion points: Reads session logs and existing agent definitions from ~/.claude/agents/.- Boundary markers: Employs structured task logic and requires explicit user approval before persisting changes.- Capability inventory: Uses Read, Write, Bash, and Agent tools necessary for ecosystem management.- Sanitization: Includes a dedicated 'Secret scan' step in the meta-distill pattern to prevent the leakage of API tokens, OAuth secrets, or sensitive environment variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 08:04 PM
Security Audit — agent-trust-hub — meta-prompting-suite