菜单
Audit the dissertation agent's rule stack, memory layers, skill routing, tool discipline, window separation, persistence, and output gates for conflicts, stale context, hidden assumptions, or maintenance risk.
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
Design and review active learning implementation support for university teachers, including classroom activity structures, preparation workflow, student participation, assessment alignment, and AI-agent support points.
Turn dissertation evidence into an AI agent design specification for supporting active learning implementation, including user scenarios, functions, interaction flow, boundaries, failure modes, and design rationale.
Synthesize confirmed co-design, design-elicitation, prototype-feedback, or participant-generated design outputs into requirements, design principles, prototype changes, participant contribution maps, and design decision logs.
Design specifications for dissertation figures such as conceptual frameworks, concept-card/interview process diagrams, AI-agent workflow diagrams, adoption-condition models, and findings maps.
Maintain a lightweight dissertation research wiki with project memory, literature notes, methodology decisions, supervisor feedback, design rationale, and open questions.
Audit an AI agent prototype or design concept for usability, pedagogical fit, teacher feedback, active learning support, evidence quality, and next design iterations.
| name | dissertation-agent-architecture-audit |
| description | Audit the dissertation agent's rule stack, memory layers, skill routing, tool discipline, window separation, persistence, and output gates for conflicts, stale context, hidden assumptions, or maintenance risk. |
Use this skill for deeper maintenance audits of the dissertation agent system, especially when rules conflict, behavior degrades, the user reports unreliable outputs, or new skills/workflows are added.
Adapt ECC agent-architecture audit to a local dissertation workflow. This is not a software production audit. It checks whether the current rules, skills, memory files, and window roles are coherent and source-grounded.
| Layer | Dissertation Version | Common Failure |
|---|---|---|
| 1 | System/developer instructions | conflicts with project rules or window role |
| 2 | Current chat history | old task dominates current request |
| 3 | AGENTS.md | skill list or workflow outdated |
| 4 | PROJECT_AGENT_PREFERENCES.md | preferences duplicate or conflict with skills |
| 5 | .agents/skills/ | overlapping triggers, missing frontmatter, stale rules |
| 6 | research-wiki/TASK_STATE.md | stale next action or missing handoff |
| 7 | knowledge-base / Obsidian | duplicate facts, broken links, unclear source of record |
| 8 | tool surface | browser, Word renderer, LMS, Obsidian, connectors assumed but not verified |
| 9 | output style rules | final answer shape conflicts with task need |
| 10 | persistence | local files, memory notes, and generated documents disagree |
AGENTS.mdPROJECT_AGENT_PREFERENCES.mdresearch-wiki/TASK_STATE.mdresearch-wiki/WINDOW_WORKFLOW_PROMPTS.mdSKILL.md filescritical: can produce wrong formal or ethical outputhigh: likely to cause repeated workflow failuremedium: creates confusion or duplicated effortlow: minor maintainability issueTASK_STATE.md after confirmed rule changes.Finding:
- Severity:
- Layer:
- Evidence:
- Risk:
- Fix:
- Affects Production Window: yes/no
agent-orchestration.