Nishi vs the Field -- Autonomous Team and Software Agents

Autonomous Team and Software Agents — mechanically measured, liar-killed, sovereign.

Capability	Nishi	SWE-agent	OpenHands	Aider	AutoGen	Notes
Autonomous execute-and-verify loop on machine targets	Yes	Yes	Best	Yes	Yes	LIVE-gated 2026-07-06: 3/3 targets advanced (2 evo-synth held-out + compiler-health witness); its own output prints the honest coverage: 0 of 104 stalled workstreams declare a machine-executable target yet
Anti-drift coach gating rung claims (liar-killer)	Yes	No	No	No	No	Gate GREEN then census then coach; only COACH-OK claims a rung -- checks U-axes, dated bars, claim-vs-proof ratio, freshness; no field agent gates its own success claims against measured censuses
Work board with heartbeat liveness	Yes	No	Part	No	Part	The board refuses to trust a stale ACTIVE label (imports nx_heartbeat_str, wellformed-checked); OpenHands and AutoGen track task state without staleness refusal
Governed scheduler (warden-owned conductor pulses)	Yes	No	No	No	Best	Ratchet and training pulses run ONLY governed, never ad-hoc; AutoGen is the orchestration-framework bar
Organ lineage genealogist	Yes	No	No	No	No	Ancestry and relatedness of organs tracked as data; the field tracks conversation history, not artifact lineage
Honest autonomy metering	Yes	No	No	No	No	Measures the Claude-touch bar toward 0 -- the team scores its OWN dependence honestly; no field agent measures its own autonomy
Teacher and practice loop	Yes	No	Part	No	Part	Toy-practice honest (the 07-05 diagnosis); OpenHands and AutoGen have tutorial-ish agent training flows; NOTE this organ is one of the 19 live shadow-name collisions (runtime + _hdl_build copies differ)
LLM-driven code authoring (novel organ building)	No	Yes	Best	Yes	Yes	THE headline absent: every novel organ falls to Claude (diagnosis 2026-07-05, unchanged today); LLM code-authorship is the field's entire product
Issue-to-patch repository pipeline	No	Best	Yes	Yes	No	SWE-agent was born from SWE-bench issue resolution
Git-native pair-programming edit loop	No	Yes	Yes	Best	No	Aider is the bar: repo-map + diff edits + auto-commit
Sandboxed per-task agent runtime	No	Part	Best	No	Part	OpenHands docker runtime is the bar; our container host exists as a SEPARATE domain (950/1000) but is not wired under the team
Multi-agent conversation framework	No	No	Part	No	Best	AutoGen is the bar; our conductor schedules pulses, it does not orchestrate agent dialogue
Long-horizon planning and memory for agents	No	Part	Part	Part	Part	Hot frontier axis; our WMS registry and memory files serve Claude, not the team organs
Benchmark-scored coding evaluation	No	Best	Part	Part	Part	SWE-bench pass-rate is how the field proves itself; we have qabench for research, nothing scores team coding
Human-in-the-loop approval flows	No	Part	Yes	Yes	Part	Operator gating exists by DOCTRINE and the workflow domain ships DENY-WINS approvals, but no team organ carries approval flows
Web UI for agent sessions	No	No	Best	No	Part	OpenHands app is the bar; our /project portal shows projects, not live agent sessions
Cost and token budget management	No	Part	Part	Part	Part	Every field agent meters provider spend; ours has nothing to meter (see the zero-token exceed) -- absent as a feature, moot by construction
Zero-LLM-token sovereign autonomous loop	Best	No	No	No	No	The whole execute-verify-advance loop is compiled sovereign organs: 0 tokens, 0 API spend, deterministic, offline; every field agent burns provider tokens per action
Refusal-over-fabrication by construction	Best	No	No	No	No	Gate-proven neg-control: a non-zero exit NEVER advances -- novel work is refused (NEEDS_TUTOR), not faked; the field's agents are notorious for claiming unverified success

Honest verdict. The Nishi team is a REAL governance-and-verification shell around autonomous work -- a zero-token machine execution loop that VERIFIES before advancing and REFUSES to fake novel work (gate-proven neg-control), an anti-drift coach that liar-kills SOTA claims against live censuses, heartbeat-liveness boards, governed conductor pulses, lineage tracking, and an honest autonomy meter. What it is NOT yet is a builder: the 2026-07-05 diagnosis stands unchanged -- zero of the 104 stalled workstreams declare a machine-executable target, the synthesizer covers toy arithmetic, and every novel organ still falls to Claude, while SWE-agent, OpenHands, and Aider ship LLM code-authorship as their entire product. The honest climb: give real workstreams machine-readable targets with heartbeat-on-advance, grow the synthesizer past toys, then wire an LLM-authoring lane UNDER the coach and verify gates the shell already provides -- the governance the field lacks wrapped around the authorship we lack.

Generated by nx_swcompare_matrix (sovereign NishiLang organ) from knowledge/compare/team.matrix — every Nishi cell verified against real organ source on disk. Zero JS, zero trackers.