This index describes the current repository state. Guides and references below are maintained against main; ADRs, dated plans, audits, reviews, and bot-run bilans are point-in-time records and may intentionally describe an older state.
Page
Purpose
current-state.md
Living as-built overview: shipped surfaces, runtime/backend status, security defaults, cloud control plane, and known limits.
why-iterion.md
Product rationale, workflow patterns, and the asymptote lens.
install.md
Install through the CLI, studio, desktop, Docker, cloud, dispatcher, scheduler, or TypeScript SDK.
examples.md
Maintained bot catalogue and focused DSL examples.
cli-reference.md
Complete top-level CLI map plus the commonly used commands and flags.
visual-editor.md
Browser-based studio, graph editor, launch forms, and live diagnostics.
skill.md
Install Iterion guidance into AI coding agents.
For the architectural trade-off against prompt-only orchestration, read why-not-prompt-orchestration.md . asymptote-bench.md and thinking-metrics.md cover workflow-quality measurement, and improvement-ratchet.md names what carries a gain from one run to the next.
Language and graph construction
Page
Topic
dsl.md
Language guide and map of every declaration, node family, edge form, and workflow control.
references/dsl-grammar.md
Readable grammar derived from the parser surface.
grammar/iterion_v1.ebnf
Formal EBNF counterpart.
grammar/V1_SCOPE.md
Living boundary of the additively evolved V1 grammar and AST.
references/diagnostics.md
Authoritative sparse catalogue: DSL C001–C199 plus async C240–C242, and bundle checks C200–C234.
routers.md
Five routing modes, per-item fan-out, and convergence.
groups-iteration-subbots.md
group/use, edge foreach, fan_out_each, resources, and nested bots.
human-in-the-loop.md
Human nodes and all six interaction values, including the node-specific none and async behavior.
cursors.md
Prompt-calibration cursor declarations and node activation.
supervisors.md
Concurrent run watchers and steering messages.
dsl-totality-and-tc.md
Language totality, fuel, liveness, and Turing-completeness boundaries.
Inputs, capabilities, and reuse
Page
Topic
recipes.md
In-source presets and external recipe overlays.
attachments.md
File and image inputs.
bundles.md
Deterministic .botz packaging with skills and resources.
import.md
Lossy, non-executing import of Claude Code workflow JavaScript into draft .bot files.
backends.md
Backend/model matrix, harness behavior, and provider support.
delegation.md
Choosing in-process model: calls or delegated CLI backend: execution.
permissions.md
Workflow/node permission modes and allow/ask/deny rules.
skills-library.md
Local project/global skill library and DSL skills: references.
plugins.md
Local and cloud plugins, built-ins, MCP injection, and git-backed org sources.
memory-and-knowledge.md
Memory scopes, visibility, lifecycle, CLI, and cloud quotas.
web-search.md
Tiered search/fetch/browser capabilities.
ultracode.md
Output-compression modes and backend limits.
secrets.md
Local sealed secret store and CLI workflow.
secrets-reference.md
Unified map of local/cloud secret kinds and sealing boundaries.
privacy_filter.md
Built-in PII redaction and restoration tools.
Page
Topic
bot-invocations.md
Manifest-driven command, board, schedule, and forge invocation modes.
resume.md
Current resume states, checkpoint semantics, overrides, and stale-run safeguards.
merge-policy.md
Worktree finalization, branch ownership, and merge authority.
review-merge-gate.md
Review-environment conversation and final merge gate.
sandbox.md
Docker, Podman, and Kubernetes isolation, bot/repository devbox.json tool provisioning, and egress proxy policy.
scheduling.md
Cron schedules, sub-minute keepalive, overlap guards, and audit history.
dispatcher.md
Tracker polling, leases, retries, hooks, and per-issue bot dispatch.
native-tracker.md
File-backed kanban tracker used by the dispatcher and studio.
session-board.md
Session and pipeline board projections.
repo-scope.md
Project root and repository-scope behavior.
settings-precedence.md
CLI, environment, project, user, and workflow precedence.
environment-variables.md
Reference of the ITERION_* environment variables and their effects.
config-share.md
Scoped, role-aware configuration sharing/editor surface.
browser-pane.md
Studio browser pane and isolation boundaries.
post-mortem-shell.md
Controlled shell access after a run.
persisted-formats.md
Filesystem run, checkpoint, event, artifact, interaction, attachment, plan, tool-blob, and message contracts.
observability/README.md
Prometheus, OTLP, Grafana, and operational metrics.
Bots and security automation
Cloud / agent control plane
Start with the Iterion Cloud overview for the event → queued run → result-posted-back loop, or cloud.md for the deployable components.
Users, teams, and integrations
Page
Topic
cloud-user.md
Login, teams, invitations, credentials, PATs, and password reset.
forge-integrations.md
GitHub/GitLab/Forgejo connections, app/token setup, and bot enablement.
forge-permissions.md
Least-privilege forge permissions and minted-token scopes.
forge-conversations.md
Command/reply routing for PR/MR conversations.
webhooks.md
Inbound provider/generic webhooks, authentication, idempotency, and CRUD.
outbound-callbacks.md
Signed run-result callbacks to launchers.
byok.md
Bring-your-own LLM API keys.
oauth-forfait.md
Delegated subscription/OAuth credentials.
quotas-and-limits.md
Run, cost, concurrency, rate, memory, and storage limits.
cloud-cli.md
Full iterion remote operator/user CLI.
cloud-rest-api.md
REST surface by domain and authorization class.
Operators and architecture
Architecture and contribution
Page
Topic
architecture.md
End-to-end compiler, runtime, backend, persistence, control-plane, automation, and extension architecture.
development.md
Reproducible toolchain, task graph, tests, and repository structure.
e2e_coverage.md
Stubbed end-to-end coverage map.
live-e2e-coverage.md
Credentialed/live E2E coverage and compile guards.
These collections are valuable evidence, but they do not override current code or the living references above: