From ai-writing-plugin
中文优先指导 technical_solution document type work,同时保留 architecture、performance、security、cost、rollout、source tier、provenance、sample/reference、HITL 和 no-HARA-leakage boundaries。
How this skill is triggered — by the user, by Claude, or both
Slash command
/ai-writing-plugin:technical_solutionThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill for `task_type: technical_solution` / technical solution / 技术方案.
Use this skill for task_type: technical_solution / technical solution / 技术方案.
默认用中文解释技术方案 workflow、输入材料、架构决策边界、open confirmations、review/verify 结果和最终交付物。保留 technical_solution、architecture decision、performance target、security boundary、HITL、NEEDS_USER_CONFIRMATION 等英文关键术语。
如果用户材料是英文,可以保留原文术语和引用片段;面向用户的步骤说明和风险提示优先中文。不要把中文输出写成架构批准、性能保证、安全批准、最终成本或 rollout risk acceptance。
technical_solution supports technical方案, backend, architecture, implementation, and technical review writing workflows. It organizes project context, requirements, architecture overview, data flow, interfaces, implementation plan, risks, rollout, and open issues.
It does not automatically approve architecture decisions.
technical_solution is an official L3 built-in document type backed by built-in DocumentTypeRules, fixtures, regression tests, and this domain guideline.
The command layer remains generic. There is no technical_solution-specific pipeline.
The demo fixture shows the expected material roles:
system_context.md as sourcerequirements.md as sourcesolution_template.md as templatechecklist.md as checklistarchitecture_reference.md as referencesample_solution.md as sampleThe technical_solution default sections are:
technical_solution critical claims include:
These claims need source evidence or HITL confirmation. If evidence or user confirmation is missing, keep NEEDS_USER_CONFIRMATION or list an open confirmation.
The following must not be finalized automatically:
The plugin must not automatically approve architecture decisions, guarantee performance, define final cost, approve security boundary, or accept rollout risk.
Without sufficient T0/T1 support and explicit HITL, do not write final claims such as:
These are warning examples, not recommended output.
sample_solution must not be used as fact source. sample is not fact source. reference cannot prove project facts. reference must not prove project-specific requirements or constraints. T3/T4/T5 cannot support critical claim by themselves.
Review focus includes:
Verification focus includes:
technical_solution output must not leak HARA-only terminology. The warning terms include:
These terms may appear in this Skill only as a no-HARA-leakage warning. They should not appear as technical_solution output content unless the user explicitly provides a project source that truly requires discussing HARA terminology, and even then it should be treated as a source-specific fact rather than generic technical_solution boilerplate.
The final package is for human review, not architecture approval. Valid final status values are ready_for_human_review, finalized_with_open_items, and blocked_pending_confirmation.
Unconfirmed critical claims remain pending until real HITL decisions are recorded. final report is not approval.
.venv/bin/python -m ai_writing_plugin write-run --task examples/technical_solution_demo_fixture/task.yaml
candidate_profile_update.yaml and candidate_skill_patch.md must not overwrite a stable skill.Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub exquisite0828/ai_writer_plugin --plugin ai-writing-plugin