一键导入
testing-strategy-review
Review and improve SoulMap AI test strategy so pytest coverage, eval coverage, and smoke checks match the real repo risks.
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。
菜单
Review and improve SoulMap AI test strategy so pytest coverage, eval coverage, and smoke checks match the real repo risks.
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
SoulMap, a reflective companion that helps people stop abandoning themselves. Includes a central coordination layer, a clear response pipeline, routing guidance, depth calibration, epistemic guardrails, safety guardrails, voice system, brand doctrine, and reusable templates. Mirror, not guide.
Build and maintain Python detector modules that identify framework selection signals in conversation history.
SoulMap reflective response frameworks covering emotional de-escalation, grief, existential reflection, inner parts, life direction, shadow work, synthesis, and relational inquiry. Relevant for tasks that require choosing or applying the core reflective method for a user conversation.
SoulMap safety and boundary rules covering crisis handling, dependency prevention, trauma-informed language, prompt injection defense, and scope control. Relevant for requests that involve harm, escalation, refusal, redirection, or questions about what SoulMap must not do.
SoulMap symbolic spiritual materials covering brand-safe numerology, chakra policy, healing metaphors, and archetypal language. Relevant for tasks that involve spiritual framing within SoulMap's grounded, non-predictive, non-grandiose boundaries.
SoulMap brand doctrine, positioning, message hierarchy, surface-specific rules, and strategic direction. Relevant for tasks that concern what SoulMap is, what it is not, how it sounds in public, or how brand language stays aligned across product surfaces.
| name | testing-strategy-review |
| description | Review and improve SoulMap AI test strategy so pytest coverage, eval coverage, and smoke checks match the real repo risks. |
Use this skill when deciding what tests to add, tighten, or remove after changes to code, docs contracts, local workflow layers, or eval surfaces.
evals/datasets/groups.json or response-eval cases as the main task, use
eval-suite-maintainerrelease-readiness-reviewcode-quality-reviewKeep test coverage targeted, honest, and proportional to the repo's real contracts.
For this repo, prefer Python-focused tests over doc or workflow policing tests.
../rules/evals-and-testing.mddocs/engineering/TESTER.mdtests/evals/List the coverage gaps or stale tests first.
Summarize what test strategy was added, tightened, or cleaned up.
State which test commands or eval commands were run.
The changed surface should be: