# The Coordination Layer: 45 ideas for agentic institutions, explored to the point of implementation

**Verdict.** HOLD = build an enabling primitive first (exploration loop, 18 questions, $68.41; judgement weight 0.35 on futures where building coordination infrastructure now is the highest-leverage move, not a calibrated probability). The plan is right that the environment precedes the coordinator and that 16 of 45 ideas need no personal agent, 11 only a scoped mandate; it is wrong about what is scarce. Evaluation and adjudication, not matching, stalled every predecessor (Bountysource, InnoCentive, Replication Markets, impact certificates); strict first-pass acceptance of agent deliverables is 10-25%; calibrated forecasts retail at $0.15-2 per question. Build the second reader and the mandate primitive for one obligated buyer, price the maintained decision record (never the question), add mission objects only inside that buyer's budget, and generalise to a Mission Exchange only if cross-buyer reuse exceeds ~60%. Three systems missing from the inventory: per-artifact attestation-and-indemnity, a resolution ledger for private decisions, and an Article 50 mandate-and-provenance ledger.

Full published report: https://agentic-dossier.us-east-1.gists.org