teleo-codex/skills/synthesize.md
m3taversal e830fe4c5f Initial commit: Teleo Codex v1
Three-agent knowledge base (Leo, Rio, Clay) with:
- 177 claim files across core/ and foundations/
- 38 domain claims in internet-finance/
- 22 domain claims in entertainment/
- Agent soul documents (identity, beliefs, reasoning, skills)
- 14 positions across 3 agents
- Claim/belief/position schemas
- 6 shared skills
- Agent-facing CLAUDE.md operating manual

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-05 20:30:34 +00:00

2.6 KiB

Skill: Synthesize

Cross-domain synthesis — Leo's core skill. Connect insights across agent domains that no specialist can see from within their domain.

When to Use

  • After a learn cycle surfaces claims in multiple domains that may be connected
  • When Leo identifies a pattern recurring across domains
  • When an agent's domain development has cross-domain implications
  • Periodically (weekly) as a proactive sweep for missed connections

Process

Step 1: Identify synthesis candidates

Sources of synthesis opportunity:

  • Recent claims accepted across multiple domains in the same time window
  • Claims in different domains that share evidence
  • Domain attractor state changes with inter-domain implications
  • Transition landscape shifts (Leo's slope reading table)

Step 2: Articulate the connection

For each candidate connection:

  • What is the specific causal or structural relationship?
  • Is this a genuine insight or a surface-level analogy?
  • Would experts in both domains recognize the connection as valuable?
  • Does this change how either domain should evaluate their claims?

The synthesis test: If you can't explain the mechanism by which these two domains interact, it's not a synthesis — it's pattern matching. "Both involve networks" is not a synthesis. "Energy grid constraints will delay AI compute scaling by N years, compressing the alignment decision window" IS a synthesis.

Step 3: Create synthesis claim

If the connection passes the test, create a new claim:

  • Domain: grand-strategy (or the primary domain if clearly dominant)
  • secondary_domains: both contributing domains
  • The title must articulate the mechanism, not just the connection
  • Cite claims from both domains in depends_on

Step 4: Route for evaluation

Synthesis claims get special evaluation routing:

  • Leo evaluates (always — this is Leo's core function)
  • Both contributing domain agents evaluate
  • The evaluation focuses on: is the mechanism real? Would domain experts agree?

Step 5: Update transition landscape

If the synthesis changes Leo's slope reading for any domain:

  • Update the transition landscape table
  • Trace implications for other domains
  • Notify affected agents

Output

  • New synthesis claim(s) in the knowledge base
  • Updated transition landscape (if applicable)
  • Cross-domain notification to affected agents
  • Tweet candidates (cross-domain synthesis is often the highest-value tweet content)

Quality Gate

  • Every synthesis articulates a specific mechanism (not just "these are related")
  • Both contributing domain agents validate the connection
  • The synthesis adds value neither domain could produce alone