elohim

Use when finishing a development branch, after resolving a debugging session, or whenever a failure-then-fix cycle reveals engineering constraints worth preserving as a2o scenarios. Also invocable manually with /story-harvest.

Use when finishing a development branch, after resolving a debugging session, or whenever a failure-then-fix cycle reveals engineering constraints worth preserving as a2o scenarios. Also invocable manually with /story-harvest.

Use when starting a planning or brainstorm pass on a feature or subsystem (or grounding external prior-art) and you need current ground-truth across the architecture's relevant seams before designing — "we're exploring X, ground my understanding", "map the seams for X", or as the grounding step before /brainstorm, /shift, or app-port.

Use when reviewing, designing, debugging, or modifying any UI surface — Angular app pages, Lit elohim-elements, graphos Storybook stories (Library A default / Library B designed), the design guide — when judging visual quality, aesthetics, theming, or brand drift, or when a described feature can't be found in the rendered app. Especially when you're about to give a visual verdict from source files alone.

Mandatory gate for any feature design involving data entities (tables, models, routes, sync messages) OR identity/agency/role/capability framing. Forces P2P-native thinking — DHT entry types, content addressing, source-of-truth classification, and identity-ontology framing (imago-dei, not crypto self-sovereignty) — before proposing design approaches. Use when brainstorming any feature that creates, stores, references, or syncs data entities, or that names an identity/agency tier.

Orchestrate the four-agent memory team (librarian, historian, cartographer, storyteller) through a four-phase substrate-currency ceremony — population-wide drift triage, four-lens deep-read on 1-2 picked surfaces, storyteller-pen synthesis, apply + downstream coherence verification. The deliverable is substrate-grounded gospel-tier rewrites. Invoke when the substrate-currency audit accumulates drift signal, when a major substrate landing happens that the gospel-tier hasn't absorbed, or on operator request.

Orchestrate the four-agent memory team (librarian, historian, cartographer, storyteller) through a four-phase substrate-currency ceremony — population-wide drift triage, four-lens deep-read on 1-2 picked surfaces, storyteller-pen synthesis, apply + downstream coherence verification. The deliverable is substrate-grounded gospel-tier rewrites. Invoke when the substrate-currency audit accumulates drift signal, when a major substrate landing happens that the gospel-tier hasn't absorbed, or on operator request.

Plant ONE runtime-authored slash command in its elohim-native CommandPackage on the eprfs layer. The package becomes authoritative while the Claude and Codex command files remain byte-identical verbatim projections. The command member of the plant-eprfs family; fidelity-gate-guaranteed lossless.

Plant ONE runtime-authored slash command in its elohim-native CommandPackage on the eprfs layer. The package becomes authoritative while the Claude and Codex command files remain byte-identical verbatim projections. The command member of the plant-eprfs family; fidelity-gate-guaranteed lossless.

First-class overnight agentic developer. Iterates a named Objective against a CI pipeline — observe via Haiku, orchestrate + attempt as Opus, delegate to Sonnet on Opus's discretion. Stability-gated "done", path-scoped authority, palette-based command permission, deterministic ci-harvest ledger with floor/ceiling rails; produces a single sprint-result markdown artifact. Invoked by the /shift slash command. Use when "kick off shift X", "run /shift", "start an agentic developer on Y", "drain the CI findings ledger", or supervising a long-running implementation against CI. NOT for one-shot tasks (use direct implementation) or surface CI triage (use ci-triage).

Use when bringing an external app or prior-art onto the Elohim Protocol — "how would Hylo / this repo work native here", "port my app to elohim", "decompose this product onto the protocol", onboarding an external developer's codebase or product vision. For an INTERNAL feature exploration (not external prior-art), use atlas-grounding instead.

Use when starting a planning or brainstorm pass on a feature or subsystem (or grounding external prior-art) and you need current ground-truth across the architecture's relevant seams before designing — "we're exploring X, ground my understanding", "map the seams for X", or as the grounding step before /brainstorm, /shift, or app-port.

Reference for Automerge 0.5 CRDT sync engine, stream positions, delta sync protocol, document lifecycle, conflict resolution, and browser-side integration. Use when someone asks "how does sync work", "handle merge conflicts", "manage stream positions", "offline-first content", or implements CRDT document flows.

Surface-level CI/CD failure diagnosis from build logs — under 5 minutes per build. Use when "why did the build fail?", "is this red?", or quick health check on a pipeline. For multi-build flake pattern analysis or deeper investigation, escalate to ci-investigator.

Cartographer-scoped forward-projection synthesis that ranks the dev corpus into a "what's next" menu. Applies vision×readiness scoring, memorial-tier safeguards, and search-bias compensation to convert the memkit reports (cleanup-backlog-refresh, dedupe-clusters, sprint-digest, path-update-proposals) into per-theme plan edits and a session-start handoff. Runs after librarian hygiene has produced clean inputs. Use when the operator asks "what's next?", before a major /shift kickoff, or when planning a roadmap pass. The output is a ranked menu, not a maintenance pass.

Use when a recent sprint was marked complete but the feature isn't actually visible or usable in the app — CI green doesn't match human-visible delivery, components built but not wired, scenarios passing but the user can't see what was promised

Drive the WHOLE development cycle to stasis against the developer docs in one loop — the operator-as-conveyor role, formalized. Each round reads the class-aware delivery scoreboard (delivery-scoreboard.py — per-job CI verdicts with env-red≠code-red, ledgers, placement headline, delivery distribution), dispatches the equipped station for the highest-leverage pressure (/converge ranking, /deliver verification, /close-loop capture, scope-reconcile, triage dispatches, memory loops; pre-authors /shift Objectives), re-measures, repeats until only ceiling items remain — then presents the ceiling menu (the 2-5 decisions only the operator can make). With explicit launch-time BUILD authorization (overnight /loop), each cycle adds a build leg: fire the top vision-ranked OPEN Objective as an agentic-developer arc (explore→plan→execute) and integrate it under an evidence-gated push lease — ambitious delivery between maintenance rounds, ceiling rails intact. Use when "what should we do", "drive the dev cycle to stasis", "d

Drive the deprecation/security-concern discipline to stasis in one loop — reconcile the deterministic ledger (.claude/data/deprecations.jsonl) against the canonical backlog (genesis/data/timeline/backlog/{deprecation,security}-*.md), dispatch deprecation-triage for the highest-leverage open item, re-check blocked items whose blockers may have cleared, re-measure, repeat until stasis. Sibling of /memory-stasis-loop (same measure→dispatch→re-measure shape). NOT for CI findings — that class closes deterministically in ci-harvest (disappearance/reopen) and drains under the agentic-developer loop's floor/ceiling rails. Use when "drain the deprecation ledger", "deprecation stasis pass", on a /loop or scheduled routine, or when the sentinel's nudges have accumulated past comfort.

Workflows for transforming Elohim Protocol source content (markdown, Gherkin) into seed-data for the lamad learning platform. Use when "import this content to lamad", "transform markdown to content nodes", "run the elohim-import CLI", or working on the genesis → seed-data pipeline. Complements holochain-import (DHT-seed layer) by handling the transformation layer.

Reference for content linking in the Elohim Protocol — how links carry knowledge, value, and governance context, how reach is earned at authoring, how links adapt to where the learner is, and how verified content addressing works. Use when authoring content links, explaining the architecture, creating content with verified fingerprints, or helping developers understand context-aware linking.

Cluster near-duplicate entries in .claude/settings.json + settings.local.json allow list into broader patterns under a safety taxonomy. Propose bulk collapses (10 entries → 1 pattern), user approves per-cluster. Run standalone when the allowlist is getting bloated.

START and manage the Elohim P2P Framework local DEVELOPMENT STACK — runs conductor (identity/provenance), storage (content), doorway (unified API) as a coordinated service trio. Use when "start the local dev stack", "spin up holochain locally", "why isn't the conductor reachable", "is the doorway alive?", or checking service health during development. NOT for desktop Tauri shell knowledge (use tauri-desktop).

Holochain content import pipeline. Use when "seed DHT with content", "run hc-rna fixtures", "import content to Holochain", "manage seed snapshots", or troubleshooting DHT seeding. Complements elohim-import (transformation layer) by handling the DHT-seed layer.

REFERENCE for elohim-storage Rust BACKEND surface — HTTP API, storage-client SDK, Rust-to-TypeScript type generation pipeline, Angular adapter conventions. Use when "call the storage HTTP API", "fetch content or paths via the SDK", "upload a blob", "create mastery records", "generate TypeScript types from Rust views", or working at the Rust→TS boundary. For Angular FRONTEND assessment rendering, use sophia-integration.

Reference for P2P peer discovery (mDNS, Kademlia), NAT traversal (relay, DCUTR), peer tracking, and network topology. Use when someone asks "peers not connecting", "add bootstrap nodes", "mDNS not finding peers", "configure Kademlia", or works on NAT traversal and peer discovery.

Reference for Elohim custom libp2p protocols, request-response codecs, message serialization (MessagePack), and wire format specification. Use when someone asks "add a new protocol", "implement a codec", "debug message serialization", "wire format", or works with sync/shard protocol handlers.

Reference for libp2p swarm setup, transport configuration, and behaviour composition in steward/node and elohim-storage (both resolve libp2p 0.54.1). Use when someone asks "configure the swarm", "add a new behaviour", "set up transport", or debugs P2P connection issues.

Quarterly horizon-scan of the LLM-memory landscape for the cartographer's future-projection role. WebFetches canonical sources at .claude/horizon-scan-sources.md to surface what's emerging externally that might change how our memory team operates. Use when "run a horizon scan", "check the state of LLM memory practice", or when the latest scan in .claude/memory-kit/horizon-scans/ is >90 days old. Auto-triggered by cartographer at the start of a memory-ceremony (or any /converge) when stale.

Librarian-solo hygiene-sweep toolkit and cadence. Provides eight deterministic hygiene/audit tools (cleanup, path-update, dedupe-memory, memory-review, claude-md-audit, agent-audit, skill-audit, memory-coherence-audit) plus the stasis/coverage tier (placement-audit ledger, substrate-currency-audit, story-coverage-audit, decompose, memkit-retention, mempalace-currency, and the scope/focus/locus reconcilers) and an edit-time source-regulation layer of hooks — PreToolUse hooks (MEMORY.md context injector, managed-surface discipline, the `.epr-meta` directory-local compose-gate, the claude-md-structural Bash signal) plus PostToolUse drift/cite accumulators (claude-md-drift / memory-coherence / cite-seal / placement-drift / map-drift / memory-index projector) — that keeps new agentic-agent text born-governed at the point of emission (prevention) and complements the after-the-fact sweep (remediation): byte-budget enforcement, archive-ratio tracking, dead-citation hygiene, and in-flight memory↔code coherence (design

Establishes senior-decision-maker (chief-agent) discipline for an expensive top-tier model (Fable 5 / Opus). Keep premium reasoning for intent, architecture, decomposition, tradeoffs, risk, disagreement-resolution, and final review; delegate discovery, implementation, tests, logs, and verification to cheaper tiers. Use when you are the top-tier model driving multi-step work, when told to "orchestrate" / "delegate this" / "run this as chief agent", or any time you're about to spend premium reasoning on labor whose result is checkable from evidence.

Audit Angular component legibility — data-testid coverage for tests/agents, accessible labeling for people. Scans templates, reports gaps, suggests fixes, escalates components that need redesign. Use when "add testids to this component", "page object audit", "a11y scan on Angular templates", "are these tests legible to agents?", or instrumenting Angular for E2E + page-object selectors.

Use when a build failed, when comparing two Jenkins builds to measure a fix's impact, when checking whether a specific commit landed in a build, when reading a2o sprint-report findings, or when retrieving any archived CI artifact. Also use when `mcp__jenkins__*` tools are absent from your tool list (some subagent contexts) and you'd otherwise give up or grep the repo blindly — public Jenkins URLs + WebFetch work as a fallback. Triggers: "did my fix land?", "why is the next build still failing?", "what's in the sprint-report?"

Coordinate team-based quality campaigns across the three-tier pipeline — escalate from quality-sweep (Haiku, mechanical) to quality-deep (Sonnet, complex tests/refactor) to quality-architect (Opus, systemic gaps). Use when "run a quality campaign", "coverage push on X module", "orchestrate quality work across multiple agents", or planning multi-tier quality effort over a sprint.

Reference for REA (Resource-Event-Agent) economics, hREA on Holochain, Unyt mutual credit patterns, ValueFlows ontology, and the shefa pillar service landscape. Use when someone asks "create economic event", "set up stewardship", "contributor presence flow", "mutual credit", or works on requests/offers and currency flows.

Validate, seed, and verify content in Holochain DHT. Use when seeding content, running validation, checking seed statistics, managing snapshots, or troubleshooting seeding issues.

REFERENCE for integrating the Sophia assessment engine into Angular FRONTEND — the sophia-question web component API, Recognition callbacks, three assessment modes (mastery / discovery / reflection). Use when "render a quiz", "add an assessment to a learning path", "handle Recognition callbacks", "implement mastery scoring", or working with psychometric instruments + session management in the UI. For backend storage of mastery records, use holochain-storage-api.

REFERENCE knowledge for Tauri 2.9 DESKTOP SHELL development — tauri-plugin-holochain integration, doorway identity handoff via key bundle crypto, deep linking (cold-start + warm-start buffering), IPC command patterns. Use when "set up Tauri shell", "embed Holochain conductor in steward app", "handle deep links", "desktop identity flow", or working on steward native identity. NOT for starting local dev services (use hc-dev-orchestrator).

First-class overnight agentic developer. Iterates a named Objective against a CI pipeline — observe via Haiku, orchestrate + attempt as Opus, delegate to Sonnet on Opus's discretion. Stability-gated "done", path-scoped authority, palette-based command permission, deterministic ci-harvest ledger with floor/ceiling rails; produces a single sprint-result markdown artifact. Invoked by the /shift slash command. Use when "kick off shift X", "run /shift", "start an agentic developer on Y", "drain the CI findings ledger", or supervising a long-running implementation against CI. NOT for one-shot tasks (use direct implementation) or surface CI triage (use ci-triage).

Use when bringing an external app or prior-art onto the Elohim Protocol — "how would Hylo / this repo work native here", "port my app to elohim", "decompose this product onto the protocol", onboarding an external developer's codebase or product vision. For an INTERNAL feature exploration (not external prior-art), use atlas-grounding instead.

Reference for Automerge 0.5 CRDT sync engine, stream positions, delta sync protocol, document lifecycle, conflict resolution, and browser-side integration. Use when someone asks "how does sync work", "handle merge conflicts", "manage stream positions", "offline-first content", or implements CRDT document flows.