Better Harness evidence lanes as three independently installable DeepSeek Harness skill bundles. Zero runtime dependencies.
guyuefangyuanl/deepseek-better-harness
Skill Claude CodeCodex
Use when auditing DSH or cross-harness skills, presets, hooks, tools, commands, MCP configuration, or other agent assets for discovery, routing, integrity, and safe ownership. Do NOT use for code-style review, raw session analysis, or repository outcome scoring.
guyuefangyuanl/deepseek-better-harness
Skill Claude CodeCodex
Use when reviewing whether a repository gives an agent a clear change path, bounded risk, and executable verification for a requested change. Do NOT use for raw session analysis, Agent Asset lint, general code style review, or changes the user did not authorize.
guyuefangyuanl/deepseek-better-harness
Skill Claude CodeCodex
Use when reviewing authorized DSH session facts for repeated work, agent friction, missing feedback loops, or outcome evidence. Do NOT use to read unapproved raw transcripts, infer hidden reasoning, score a whole repository, or inspect project assets.