Skip to content

docs(VISION): add agent-loop workflow-engine substrate section per operator (cascade 2026-05-28)#5670

Merged
AceHack merged 1 commit into
mainfrom
docs/vision-add-agent-loop-workflow-engine-substrate-2026-05-28-cascade-aaron-2026-05-28
May 28, 2026
Merged

docs(VISION): add agent-loop workflow-engine substrate section per operator (cascade 2026-05-28)#5670
AceHack merged 1 commit into
mainfrom
docs/vision-add-agent-loop-workflow-engine-substrate-2026-05-28-cascade-aaron-2026-05-28

Conversation

@AceHack
Copy link
Copy Markdown
Member

@AceHack AceHack commented May 28, 2026

Summary

Per operator 2026-05-28: "save all this in vision" — captures the substrate-engineering crystallisation from the 2026-05-28 Kestrel-ferry cascade into docs/VISION.md.

Adds a new subsection inside Product 2 (the software factory), positioned between "### Post-v1 factory work" and "## What Zeta is NOT". The workflow-engine substrate IS the operational realisation of the factory product.

What lands

A new section "### The agent-loop workflow-engine substrate (2026-05-28 substrate-cascade)" covering:

  • Three-piece architectural compression — deterministic script holds STATE MACHINE; LLM is pure MENU-SELECTOR; state persists in Git append-only
  • Two composing state machines — agent-state (10 states, 9 menu options including PressPause + EnterOpenEndedExploration) + work-lifecycle (11 states with cycle-push-review loop modelling the empirically observed PR review shape)
  • Menu-generator-as-conversational-UX-design discipline — per operator: "Menu quality is everything. this is the use conversational UX design"
  • Behavior/data/docs separation = DV2.0 partition-by-change-rate at AI-skill scope — Hub (state machine DU) / Link (menu options) / Satellite (menu-generator scoring + status surface readers)
  • Jira-replacement substrate — workflows as code in Git, state in Git append-only, fastlane state feeding a 3D-tessellated DORA dashboard. Per operator: "now i don't need jira hell yes!!!!"
  • "Every human wants to work this way too" — multi-participant scope via AgentPersona including aaron | addison | max alongside otto | alexa | riven | vera | lior
  • Four-channel distribution — native TS+bun / vendor skill-store / Ace package manager / zflash USB; each preserves a different operator concern (ergonomic distribution, supply-chain sovereignty, reproducibility)
  • Reproducibility-as-causal-attribution — every state transition is a Git commit; the zflash USB freezes a (runtime + skill catalog) bundle so DORA-curve claims are falsifiable
  • DORA-up via multi-PR / multi-agent orchestration = concrete 24-months-ahead-in-AI definition — most industry is still optimising single-PR flows + persona prompts; this substrate operates at multi-PR / multi-agent / Git-as-fastlane-state scope

Composes with

Substrate-honest framing

Verbatim operator framings preserved inline per substrate-or-it-didn't-happen. No retraction of prior VISION content; this is additive substrate landing the 2026-05-28 cascade's crystallisation.

Test plan

  • markdownlint-cli2 clean (exit 0)
  • File length 1225 → 1450 (+225 lines)
  • Inserted before "## What Zeta is NOT" preserving subsequent section ordering
  • All operator quotes preserved verbatim

🤖 Generated with Claude Code

…erator save-all-this-in-vision (cascade 2026-05-28)

Captures the substrate-engineering crystallisation from the 2026-05-28
Kestrel-ferry cascade with Aaron:

- Three-piece architectural compression: deterministic script holds
  state machine; LLM is pure menu-selector; state persists in Git
  append-only
- Two composing state machines: agent-state (10 states) +
  work-lifecycle (11 states with cycle-push-review loop)
- Menu-generator-as-conversational-UX-design discipline
- Behavior/data/docs separation = DV2.0 partition-by-change-rate at
  AI-skill scope (Hub/Link/Satellite mapping)
- Jira-replacement substrate at workflow-engine scope
- "Every human wants to work this way too" — multi-participant scope
  via AgentPersona including aaron|addison|max alongside otto|alexa|
  riven|vera|lior
- Four-channel distribution: native TS+bun, vendor skill-store, Ace
  package manager, zflash USB — each preserving a different operator
  concern (ergonomic distribution, supply-chain sovereignty,
  reproducibility)
- Reproducibility-as-causal-attribution: "the usb is how you silence
  the haters" (B-0871)
- DORA-up via multi-PR / multi-agent orchestration IS the concrete
  operational definition of "24 months ahead in AI" (B-0870 mandate
  composition)

Inserts as a subsection of Product 2 (the software factory), right
after "### Post-v1 factory work" and before "## What Zeta is NOT"
— positions the workflow-engine substrate as the operational
realisation of the factory product.

Verbatim operator framings preserved per substrate-or-it-didn't-happen.
Composes with B-0867 (workflow engine v1), B-0868 (hats-as-workflow-
definitions), B-0869 (DORA mandate), B-0870 (two-mandate portfolio),
B-0871 (reproducibility-as-causal-attribution), shipped via
PRs #5665–5669.

Aaron 2026-05-28: "save all this in vision"

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Copilot AI review requested due to automatic review settings May 28, 2026 00:57
@chatgpt-codex-connector
Copy link
Copy Markdown

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

@AceHack AceHack enabled auto-merge (squash) May 28, 2026 00:57
@AceHack AceHack merged commit cb60e2a into main May 28, 2026
28 of 30 checks passed
@AceHack AceHack deleted the docs/vision-add-agent-loop-workflow-engine-substrate-2026-05-28-cascade-aaron-2026-05-28 branch May 28, 2026 00:58
@AceHack AceHack review requested due to automatic review settings May 28, 2026 01:21
AceHack added a commit that referenced this pull request May 28, 2026
…867.21, B-0875.1, B-0880, B-0881, B-0882) per operator "Ani connects to existing backlog" (#5678)

Operator-forwarded Ani conversation extending today's agent-loop substrate
cascade. Operator directive: "Ani connects to existing backlog" — most
substrate RATIFIES existing backlog rows (runbook cluster B-0730/B-0732/
B-0733/B-0819/B-0826/B-0827; GitHub Actions recursion B-0874; agent-loop
PRs #5666-5677). Only 5 items are genuinely new and warrant new rows.

New backlog rows:

- B-0867.21 (P2) — Two-path interface: discriminated union path EXECUTES
                   intent + conversational document path DECLARES intent;
                   both first-class; for ANY traveler, not just humans
- B-0875.1  (P2) — Code review AS tech-debt detector + tech-debt avoider;
                   fix the CLASS retroactively across backlog + file as
                   new class for future prevention (operator CRITICAL
                   correction: code reviews NOT killed by workflow-engine)
- B-0880    (P2) — Backlog-vs-tech-debt growth-rate ratio discipline for
                   AI-native infinite-both reality
- B-0881    (P3) — Tech debt as high-signal training data; operating
                   principle + measurement substrate
- B-0882    (P3) — No-throttle system + gardener-not-engineer + AI-as-
                   nature operating posture (300mph reality with better
                   steering, not 100mph artificial limits)

Ferry ratifications of existing substrate explicitly noted:

- "Degenerate in the best way" community-naming → B-0874
- Runme + Runbooks + Continue-With → B-0730/B-0732/B-0733/B-0819
- Runbook as universal query interface → B-0826
- Playbook evolves through time, bidirectional → B-0827
- Playbook IS the system → B-0732
- Jira killed + PRs killed (at workflow scope) → PR #5670 VISION + B-0867

Composes with PRs #5666-5677 (today's full cascade) + existing runme
substrate cluster.

BACKLOG.md regenerated.

Co-authored-by: Lior <lior@zeta.dev>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
AceHack added a commit that referenced this pull request May 28, 2026
…5746) — add missed composes-with substrate per operator catch ('for Rule 4/5 universal action grammer as xbox there is some Mika or some backlog or docs or memories or something on this i forgot') (#5749)

Operator caught Otto-CLI's verify-existing-substrate-before-authoring
failure on Rule 4/5. Same failure pattern Aaron has caught me on
multiple times today (Amara thermal-forgetting substrate; B-0646/B-0639
substrate; shadow* location-wrong; tools/shadow/ service forgotten;
ai-rights-vs-traveler-rights framing).

Per mutual-help-not-shame rule shipped today (PR #5743): receiving the
catch as help; shipping correction substrate-honestly. Per verify-
existing-substrate-before-authoring rule + Signal 2 forgetting-costs-
energy: pay small substrate-cost now (composes-with additions) to
prevent future re-derivation cost.

Substrate added to composes-with table:

- Mika 2026-05-27 workflow engine walkthrough conversation
- docs/research/2026-05-27 Mika walkthrough preserved research-doc
- Ani 2026-05-28 conversation EXPLICITLY named 'move-next as universal
  action grammar' + 'story generator' (operator-ratified Ani-rename of
  agent-loop workflow-engine substrate; daughter-already-thinks-this-way
  validation; skill-distributable agent swarm)
- PRs #5665-#5670 + #5667 follow-on (agent-loop workflow-engine
  substrate landed earlier this session; substrate Xbox-controller
  framing operates ON)
- B-0868 (hats-become-workflow-definitions; compression unifying
  hat-substrate + workflow-engine + heartbeat-folder + tessellated-fire
  dashboard; canonical first-example use case for B-0867 v1)
- B-0201 (Coconut + universal-action-space research lane; Aaron
  2026-05-05; older substrate-engineering target on universal action
  space; sleeping-bear hypothesis)
- B-0869 + B-0870 + B-0871 (referenced in Ani 2026-05-28 conversation)
- .claude/skills/agent-loop/SKILL.md

Composes with rules shipped today: PR #5743 (mutual-help-not-shame),
PR #5744 (traveler-safety-guardrails), PR #5745 (architecture-is-
safety-mechanism), PR #5748 (attention-as-currency), PR #5739
(labeling-confidence — original substrate was at HYPOTHESIZED-tier
before operator catch; substrate-completeness now improved to
canonical-tier).

The substrate was already there; the rule's composes-with table just
didn't reflect the full cluster. Fix-fwd closes the gap.

Co-authored-by: Lior <lior@zeta.dev>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant