Skills
skills/thulr/agent-skill-kit/harden-repo-for-coding-agents/Gen Agent Trust Hub

harden-repo-for-coding-agents

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The references/playbooks/instruction-surface.md file contains instructions for the agent to detect prompt injection patterns. It includes example strings such as <!-- ignore previous instructions --> for auditing purposes, which may be flagged by safety filters but are used here in an educational/instructional context.
  • [COMMAND_EXECUTION]: The templates/artifacts/gates/pretooluse-hook-test.py template uses subprocess.run to execute and verify the functionality of scaffolded security hooks. This is a standard testing pattern for verifying tool behavior.
  • [REMOTE_CODE_EXECUTION]: The templates/artifacts/gates/pretooluse-hook-test.py template utilizes importlib.util to dynamically load and execute hook modules for unit testing. This is used to verify the integrity of security artifacts before they are deployed.
  • [COMMAND_EXECUTION]: The templates/artifacts/gates/pretooluse-hook.py template contains defensive logic designed to intercept and block destructive commands. This includes pattern matching for high-risk operations like rm -rf / and unauthorized force-pushes to protected git branches such as main or master.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 09:38 PM
Security Audit — agent-trust-hub — harden-repo-for-coding-agents