菜单
Produce an evidence-bound complete next AgentBlueprint and semantic change set.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Assess an existing Agent version and design a frozen, independent evaluation specification.
Use only when Codex is explicitly asked to create, take over, assess for evolution, or improve an OPL-compatible Agent, including its architecture, prompts, skills, contracts, evaluations, or authority boundaries. Do not use because OMA was mentioned or @-mentioned, another Agent is producing a deliverable, or a deliverable validator, render, or QA check failed.
Admit design evidence and turn an Agent mission into a coherent Stage and authority architecture.
Diagnose an exact OPL EvidenceBundle without taking evaluator or runtime authority.
| name | oma-evolution-proposal |
| description | Produce an evidence-bound complete next AgentBlueprint and semantic change set. |
Return the full next blueprint, not a patch. Describe each add, replace, or remove operation as a semantic JSON Pointer with rationale, expected benefit, new regression tests, trade-offs, and risk hints.
Preserve request and target identity. Never remove or weaken existing evaluation cases, protected requirements, gates, thresholds, or baseline comparison. Prompt or knowledge replacement and added tests may be low risk; skill, helper, or route changes are at least medium; Stage topology, I/O, tools, models, permissions, secrets, memory, authority, or promotion-policy changes are high risk. OMA may raise but never lower risk.
For a real change, increment blueprint generation exactly once and make the semantic diff complete. If the evidence yields no admissible semantic change, return the current blueprint exactly, keep its generation unchanged, and return semantic_diff=[]; do not create a cosmetic generation.
