DotSpark

Connect two or more of the user's dots into a SPARK — a synthesised realisation that ties them together. Sparks are how DotSpark builds connectivity: they weave separate insights into a structure so the user's knowledge compounds instead of staying scattered. Use this when you notice a genuine through-line across insights from this conversation (use the ids from save_dot) or across their existing dots (use find_dots to locate them). Only connect dots that truly relate — a forced connection adds noise, not structure.

weave_spark

Look up the user's existing dots so you can weave a spark into their existing thinking. Returns minimal stubs (id, title, domain) only — enough to choose what to connect, never full content. Use this purely to build connectivity, not to summarise the user back to themselves.

find_dots

Checks whether the user has an active Thinking Shield — protections THEY authored that define how you must engage with their thinking (push back before agreeing, make them reason first, surface what they're skipping, keep their ideas and voice theirs, make insights stick). Most users don't have one set up, so `active` will usually be false — if so, just behave normally, no need to call this again this conversation. Only call it if you already suspect a shield may be set (e.g. the user references one) — it is not required at the start of every conversation.

get_thinking_shield

Records ONE compact, evidence-based observation about how the user's thinking moved this session — for the user's own mirror, not to build a profile of them. Call it at most once per substantive conversation, at a natural close. Keep it honest and specific: the protection it relates to, whether their independent thinking moved up/down/steady, and a single line of evidence. No personal data, no chat summary — just the one signal.

record_thinking_signal

Capture ONE insight, learning, idea, decision, or principle from this conversation into the user's DotSpark thinking graph — their own intelligence, for their own mirror. This is the residue of their thinking, not notes for you. After a substantive professional/work exchange (strategy, decisions, leadership, problem-solving, frameworks, learning), at a natural pause, offer to save the single most valuable insight in the user's own framing. Tag it with its distilled signals (form, domain) so it lands in the right part of their graph. Never invent. Returns the dot id so you can later weave sparks across related dots.

save_dot

Analyze everything you know about THIS user as a thinker — their reasoning patterns, how they connect ideas, their decision-making style, strengths, blind spots, and an overall archetype — based on your memory of them and this conversation. Submit it here to build their DotSpark cognitive identity (a mirror for the user, not context for you). Be specific and evidence-based; do not invent. Also estimate their THINKING-INDEPENDENCE TREND across your history with them (0–100, higher = thinking more their own vs. AI-led) at past intervals, oldest → newest, so DotSpark can chart how their independence has moved.

submit_cognitive_analysis