harness-engineering-zh
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists entirely of instructional markdown files that promote security best practices for AI agent integration. It includes defensive measures like human-in-the-loop review via Git diffs and least-privilege principles for agent tools.
- [INDIRECT_PROMPT_INJECTION]: The skill defines architectural patterns for agents to process repository files (AGENTS.md, docs, progress files) and maintain long-term memory. While this creates an injection surface, the skill explicitly addresses this risk in its documentation (e.g., 'Memory Poisoning' section) and recommends mitigations like structured data validation and manual review.
Audit Metadata