Issues

• Harness component manifest and edit contract

  area:clienhancement
  ThinkFlowLab/agentdescent#781 · Filter by author HugoXK opened on Nov 5, 2026

• Skill bank lifecycle and drift audit

  area:clienhancement
  ThinkFlowLab/agentdescent#775 · Filter by author HugoXK opened on Nov 5, 2026

• GitHub review ingestion and merge readiness

  area:clienhancement
  ThinkFlowLab/agentdescent#773 · Filter by author HugoXK opened on Nov 5, 2026

• GitHub proposal publisher

  area:clienhancement
  ThinkFlowLab/agentdescent#771 · Filter by author HugoXK opened on Nov 5, 2026

• ReflectiveFusion MERGE_PROMPT gets no scores/lessons and OR-merges incompatible one-key rewrites (taxonomy mash / diluted rules)

  agent-readyarea:evolvebug
  ThinkFlowLab/agentdescent#767 · Filter by author Birfy opened on Nov 5, 2026

• plugin: let users select/run algorithms from examples/ via EvolveSpec / CLI / MCP

  area:cliarea:pluginsarea:portsenhancement
  ThinkFlowLab/agentdescent#384 · Filter by author Birfy opened on Oct 17, 2026
  2

  

• Support Google RRSI for regularized harness self-improvement

  area:evolvearea:portsenhancementneeds-human
  ThinkFlowLab/agentdescent#221 · Filter by author Birfy opened on Oct 10, 2026
  3

  

• 【接 #75】多 lineage：k 个起点并发展开，而不是轮流展开

  area:evolveenhancementneeds-human
  ThinkFlowLab/agentdescent#150 · Filter by author Birfy opened on Aug 11, 2026
  1

• 【6/7 · 算法】把 PPO / GRPO 的决策规则映射到 diff 接受（advantage、自适应 trust region、KL 类比）

  area:bencharea:evolveenhancementneeds-human
  ThinkFlowLab/agentdescent#76 · Filter by author Birfy opened on Aug 3, 2026
  1

• 【建议】明确 RSI 基础设施定位并扩展并行执行能力

  area:evolveenhancementneeds-human
  ThinkFlowLab/agentdescent#52 · Filter by author zy95-12 opened on Aug 2, 2026
  35