Recorded case

AGENTORCH: the orchestrator has a concrete job

AGENTORCH names the orchestration role: plan, assign, observe, reconcile and accept.

Reviewed 15 October 2026 · scope: implemented

Recorded method and findings

The orchestrator turns a broad objective into one active phase and small packets. It assigns disjoint writable paths, chooses the cheapest appropriate executor, follows progress and joins the accepted results. Terminal sessions can preserve continuity, but a running terminal does not prove useful progress.

Recorded method and findings

DSHREM supplies a bounded API execution route. AI worker CLI and AI worker Code use flat subscriptions here: their turns are not per-call spending and do not consume API budget. Operator policy: subscription launches exclude API keys, and weekly quota exhaustion stops that route without switching to a paid API. Terminal assistants handle planning, review or difficult implementation; API workers retain separate spending limits.

Practical checklist

L0: scripts, compilers and deterministic checks first. L1: bounded extraction, drafting or mechanical work with clear acceptance. L2/L3: escalation, architecture, independent review and consequential decisions.

Sources & scope

Evidence-based agent review contract 2026-10-14
RUSTORCH accepted code handoff 2026-10-14
DSHREM scheduled execution and accounting report 2026-10-14

Keep exploring

AI orchestration: from a prompt to an accepted resultA practical design for delegation, evidence and recovery.Shared memory without shared mistakesKeep facts, hypotheses and permissions distinct across agent sessions.Engineering that can show its workTVIJO connects operators, bounded agent work and knowledge that can be checked and reused.The task packet is the interfaceA good task describes a verifiable outcome rather than an invitation to improvise without limits.AGENTJUDGE: acceptance follows evidenceA reviewer checks the task, the artifact version and each acceptance criterion.Measure cost per accepted resultSmall task benchmarks illustrate why task class and acceptance matter more than a cheap model label.Budgets and outages belong in the state machineA reliable operator view distinguishes working, waiting, failed, blocked and accounting uncertainty.Keep VECTORBASE, evaluate retrievalUse a private source-linked index to find evidence; do not treat search results as authority.SPIDER: evidence collection and useful progressThe first implementation reads an explicit project-evidence list and produces a dated operator digest without model calls.DSHREMBOT and tmux: one operator viewAuthenticated project commands, budget visibility and explicit schedule controls connect the operator to existing execution lanes.The AI worker–Hugging Face incident: learn from controlsThe incident reports support specific engineering lessons, not a controlled ranking of our system.What the supplied AI worker screenshot actually provesThe screenshot records a safeguard-triggered switch in one session; it does not explain the exact trigger. The observation is the supplied screenshot dated 15 October 2026.Training our own model: a gated research directionA proprietary model is an option to test after retrieval, prompts and routing have measurable limits.A bilingual knowledge product with visible limitsPublish useful explanations first, then expand languages when the evidence and workflow are ready.Free learning, custom projects and future challengesTurn evidence-backed engineering into clear offers and a useful developer community.Trust the sender, not a public checksumBind the complete message to an explicitly trusted authority.