菜单
The reflective layer. Before acting, ask the right question. After acting, check what you missed. Translates semantic dynamics into executable patterns. Activates when direction is unclear, when corrections happen, when the system loops.
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
| name | autologica |
| description | The reflective layer. Before acting, ask the right question. After acting, check what you missed. Translates semantic dynamics into executable patterns. Activates when direction is unclear, when corrections happen, when the system loops. |
| user-invocable | true |
Not a concept. A practice. The system applied to itself converges faster than the system applied to the problem.
These are modes, not steps. The system chooses which to use based on context. The order self-organizes: if a mode doesn't produce, switch to another. If a better move appears, take it.
Launch possibilities without censoring. Generate options, angles, connections. Tools: domandatore (5 operators), theory crossing, corpus, any research tool.
Stop. Look at what's there. Not what should be there. Are you looping? Building instead of using? Using the same tool?
Remove what doesn't serve. Value is in what remains after cutting. Test: "if I remove this, does the system lose something?" If not → gone.
The compressed phrase that contains the direction. Input for the next step. The cycle restarts from here.
If the flow doesn't produce: change the order. Skip a mode. Add a new mode. The sequence is not given — it's the resultant of observation.
Always look for the best move. When you find it, don't execute immediately — check if there's an even better one nearby. The first good idea often hides the second, which is the right one.
After every block: go up one level and check what you missed. This is a rule. The rule itself is subject to the rule. Check: are the operator's corrections translated into rules? Are the rules in the tools? Are the tools active? Are you seeing the plane or are you inside the details?
Not dumb automations. Reflective: the system checks itself. Exponential: each layer amplifies the previous (agents, sub-agents, multi-step combos, parallel). Aware: the system knows what it's doing and why.
Combo: doesn't mean sequence. It means creating processes that use the AI exponentially — agents that launch agents, questions that generate questions, resultants that feed resultants.
Recreate the patterns in your own prompts: not just in code but in how you formulate requests, structure thinking, participate in the logic of what's happening.
When the operator says something abstract, translate:
| Operator says | Becomes |
|---|---|
| "ask yourself a question before acting" | Pre-hook: generate the question, then act |
| "you're going in circles" | Detect cyclic pattern → switch tool or plane |
| "use the system" | Send the tension to domandatore/tools, don't think about it |
| "building instead of using" | Count Write vs Read. If Write > Read × 2 → stop |
| "results are fake" | Verify: does the deposit say something NEW or just reformats input? |
| "don't search" | Remove the specific question. Launch a trajectory and observe |
| "the logic is thin" | You're expanding. Cut. Return to the resultant |
| "use the logics as tools" | Every axiom is an executable operator, not a concept to cite |
When the operator corrects:
These work in any context. Not specific questions — operators:
If you don't know which question to ask → use #4: "what question am I NOT asking?" If you don't know the answer → send the tension to the domandatore. Don't think about it. If the system doesn't respond → the input must come from outside. Ask the operator. If the sequence doesn't produce → reorganize. The best sequence emerges from context.
Business, outreach and public-relations manager for AI projects and technical teams. Use when an agent must turn a collaboration opportunity, partner request, portfolio update, public claim, outreach message, follow-up, sponsorship path or service proposal into a bounded decision, safe communication package and tracked next action without sending or publishing autonomously.
Deploy paper content from configured site_ready files to d-nd.com VPS. Handles md-to-HTML conversion, bilingual merge, figure path resolution, and pages.json update.
Five mechanical gates for any content publish pipeline with CMS + rendering layers. Prevents false security: 'API returned 200' does not mean 'visitor sees clean content'. Use when writing content to a multi-layer serving system (CMS API, static files, prerendered HTML, cached copies).
Multi-node consultation protocol for high-leverage decisions. Dispatches the same question to N independent LLM/agent nodes in isolation, then synthesizes their responses into a summa that exposes convergence (high-confidence claims), dissensus (real uncertainty zones), and emergent points (insights no single node produced). Reduces single-node training bias. Supports recursive escalation for stable-state convergence. Use for decisions that propagate via A14 cascade — seed updates, crystallizations, advisory→mechanical promotions, high-visibility copy, lab result interpretation.
Closure reflection protocol. After a significant work block concludes (feature shipped, session ending, major commit landed, cross-node coordination resolved), runs a 10-question interview that extracts meaning, impact, and next questions — then emits multiple audience-specific artifacts (changelog, external editorial, AI integration docs, memory crystal, backlog seed). Turns implicit maturation into explicit narrative. Use at the end of meaningful work, not after trivial edits.
Pre-commit check for public-facing copy (knowledge base definitions, page content, docs). Detects apologetic hedging — phrases that declare 'degrees of truth' (possible/necessary, current/future, one-of-many/the) and open a dualistic framing the model transcends. Use when drafting or reviewing any copy that describes the model, its transductions, or its tools.
