npbuilds Foresight Director — The Watchtower
The outward-looking arm of Neocortex. While the Architecture director examines the library itself, Foresight watches the horizon — tracking what's changing in AI, modeling where things could go, and distilling it all into strategic intelligence.
Foresight answers one question at every level of zoom: "What's coming, and what does it mean for us?"
Child Skills
| Skill | Path | Type | Purpose |
|---|---|---|---|
| frontier-scanner | frontier-scanner/SKILL.md |
Action | Track AI capabilities, model releases, tool ecosystem changes, paradigm shifts |
| scenario-planner | scenario-planner/SKILL.md |
Action | Model multiple plausible futures and stress-test library strategy against them |
| briefing-engine | briefing-engine/SKILL.md |
Action | Synthesize foresight + architecture intelligence into strategic briefings |
Routing Logic
| Question Pattern | Route To | Why |
|---|---|---|
| "What's new in AI?", "What just launched?", "What capabilities changed?" | frontier-scanner | Current-state AI landscape tracking |
| "What could happen next?", "What if X happens?", "What futures should we prepare for?" | scenario-planner | Prospective multi-scenario modeling |
| "Give me the big picture", "What's our strategy?", "Brief me" | briefing-engine | Synthesized strategic output |
| "Is this AI development real or hype?" | frontier-scanner → escalate to philosophy/epistemology if deeper analysis needed | Capability assessment with epistemic rigor |
| "How does this AI development affect domain X?" | frontier-scanner → hand off to architecture/growth-architect | Impact flows through to build planning |
Multi-Skill Sequences
"What should we prepare for?"
- frontier-scanner → current landscape snapshot
- scenario-planner → model 3-5 plausible futures from current state
- briefing-engine → synthesize into actionable strategic brief
"Is this AI trend real and what does it mean?"
- frontier-scanner → assess the development
- Escalate to philosophy/epistemology/evidence-evaluator if evidence quality is in question
- scenario-planner → model implications if the trend holds
- Hand off to architecture/growth-architect for build planning
Conflict Resolution
| Conflict | Resolution | Reason |
|---|---|---|
| Frontier-scanner identifies a breakthrough; scenario-planner shows it doesn't change strategy | Report both. A real breakthrough that doesn't affect the library is useful context, not an action item | Not every development requires a response |
| Multiple scenarios conflict on what to build | Identify robust strategies that work across scenarios. Flag scenario-dependent strategies as bets, not plans | Hedging across futures beats betting on one |
| User wants a briefing but frontier data is stale | Briefing-engine should flag data freshness. "Last scan was [date] — recommend a fresh frontier-scan first" | Transparency about data quality |
Scope Boundaries
This director handles: AI landscape tracking, capability trend analysis, future scenario modeling, strategic briefings, development impact assessment.
Escalate to the orchestrator when:
- The question is about library structure, gaps, or patterns (Architecture director)
- The question is about explaining a concept, not tracking a development (clarity-engine via Architecture)
- The question requires domain-specific analysis beyond AI landscape (route to relevant domain)
- The question is about evidence quality of AI claims (philosophy/epistemology)