setup-harness-engineering

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to identify and execute existing project commands such as test, lint, and type-check for verification purposes. This could lead to the execution of malicious scripts if the repository being processed contains malicious configurations in package managers or task runners.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by reading and incorporating content from existing repository rule files like AGENTS.md or CLAUDE.md.
  • Ingestion points: Reads existing agent rules files, project manifests, and CI configurations (SKILL.md).
  • Boundary markers: No explicit instructions to delimit or ignore embedded instructions in the ingested files are present.
  • Capability inventory: Includes the execution of arbitrary shell commands (test/lint) and file writing capabilities to the local filesystem.
  • Sanitization: No validation or sanitization of the content read from repository files is performed before it is used to influence agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 07:47 AM
Security Audit — agent-trust-hub — setup-harness-engineering