Skip to content

Iteration Plan

Purpose

Iteration Plan is the commitment artifact of the iterate loop. Its unique job is to convert the improvement backlog’s next-iteration selection and the roadmap’s current slot into one time-boxed commitment: a falsifiable goal, one Good / one Better / one Best outcome per participating workstream with owners and acceptance evidence, deterministic trade rules, and stable-ID task tables.

It is not the live tracker. The plan owns commitment membership: task rows may execute directly after user authorization. Only when tracking is explicitly selected do rows reference runtime items; an empty item ID does not require creation. The selected tracker owns live status and execution history; otherwise use compact continuation evidence alongside the plan. Scope changes route back through the plan — hand-added tracker items never silently widen the commitment.

Example

Show a worked example of this artifact
---
ddx:
  id: example.iteration-plan.depositmatch
  authoring:
    home: repo
  depends_on:
    - example.roadmap.depositmatch
    - example.improvement-backlog.depositmatch
---

# Iteration Plan

| | |
|---|---|
| **Iteration** | IT-05 |
| **Dates** | Mon 1 Jun → Fri 12 Jun 2026 |
| **Review** | Fri 12 Jun, 30-minute demo with pilot-bank operations lead |
| **Lead** | Delivery Lead |
| **Roadmap Slot** | Roadmap order 1 — CSV import validated against real pilot-bank exports |

## Iteration Goal

> By 12 Jun, a pilot-bank operations user imports a real deposit export and
> sees validated rows with actionable errors — shown live in the review demo
> against an anonymized pilot fixture, with the import test suite green.

## Committed Outcomes

Each participating workstream commits exactly one Good, one Better, and one
Best outcome. Good is the protected floor; Best drops first, then Better.
Tiers: 🟢 Good = committed for the review · 🔵 Better = first traded under
pressure · 🟣 Best = only if the iteration runs clean.

| ID | Workstream | Tier | Outcome | Owner | Governing Artifact | Acceptance Evidence |
|----|------------|------|---------|-------|--------------------|---------------------|
| WS-1-good | WS-1 | 🟢 Good | A real anonymized pilot export imports with validated rows — fixtures collected, anonymization documented, all fixture error classes handled | Priya | Improvement backlog rank 1; FEAT-001; story test plan US-003 | Fixture set committed; tests citing US-003 ACs pass; live demo at review |
| WS-1-better | WS-1 | 🔵 Better | Upload p95 latency watch item on the pilot dashboard | Priya | Improvement backlog rank 2 | Dashboard panel live with alert threshold |
| WS-1-best | WS-1 | 🟣 Best | Per-row error messages include remediation hints | Sam | FEAT-001 | Hint text reviewed by operations lead; shown in demo |

**Not participating this iteration**: WS-2 (matching starts IT-06 per
roadmap order 2), WS-3 (reporting gated on matching)

## Trade Rules

- Best drops first, then Better — the Delivery Lead calls trades and records
  them in the status report.
- WS-1-good is never traded: without it the review demo has no honest data
  and the iteration fails. A Good at risk is an escalation, not a trade.

## Tasks

The plan owns membership. This example uses an explicitly selected tracker;
existing items are referenced by ID. A row with `—` may execute directly from
the plan after authorization only when the selected runtime contract permits
direct execution; worker-required claims cannot be bypassed. An empty ID alone
does not select tracking or automatically create an item.
The selected tracker owns live status; recovery evidence remains alongside
the plan. Workstream aliases come from the roadmap registry.

### WS-1 · Data Intake — fixtures

| ID | Task | Outcome | Owner | Due | Status | Work Item |
|----|------|---------|-------|-----|--------|-----------|
| 1.1 | Collect three real pilot-bank exports under the data agreement | WS-1-good | Priya | Wed 3 Jun | Backlog | wi-231 (existing) |
| 1.2 | Script and document the anonymization pass | WS-1-good | Priya | Fri 5 Jun | Backlog | — |

### WS-1 · Data Intake — import validation

| ID | Task | Outcome | Owner | Due | Status | Work Item |
|----|------|---------|-------|-----|--------|-----------|
| 2.1 | Enumerate fixture error classes into US-003 test cases | WS-1-good | Sam | Fri 5 Jun | Backlog | — |
| 2.2 | Implement per-row validation error messages | WS-1-good | Sam | Wed 10 Jun | Backlog | — |
| 2.3 | Add remediation hints to error messages | WS-1-best | Sam | Thu 11 Jun | Backlog | — |

### WS-1 · Data Intake — pilot dashboard

| ID | Task | Outcome | Owner | Due | Status | Work Item |
|----|------|---------|-------|-----|--------|-----------|
| 3.1 | Add upload p95 latency panel with alert threshold to the pilot dashboard | WS-1-better | Priya | Tue 9 Jun | Backlog | — |

## Risks

| Risk | Impact | Response |
|------|--------|----------|
| Pilot bank delays export delivery | H | Escalate through the pilot sponsor by 3 Jun; fall back to synthetic fixtures and flag confidence in the review |
| Anonymization strips fields validation depends on | M | Review field list with operations lead before scripting |

Reference

ActivityIterate — Measure, align, and improve. Close the feedback loop back into the planning strand.
Default locationdocs/helix/06-iterate/iteration-plan-[iteration-id].md
RequiresNone
EnablesNone
InformsStatus Report
Generation prompt
Show the full generation prompt
# Iteration Plan Generation Prompt

Document one human iteration's commitment.

## Purpose

Iteration Plan is the **commitment artifact** of the iterate loop. Its unique
job is to convert the improvement backlog's next-iteration selection and the
roadmap's current slot into one time-boxed commitment: a falsifiable goal,
one Good / one Better / one Best outcome per participating workstream with
owners and acceptance evidence, deterministic trade rules, and stable-ID task
tables.

It is not the live tracker. The plan owns commitment membership: task rows
may execute directly after user authorization. Only when tracking is explicitly
selected do rows reference runtime items; an empty item ID does not require
creation. The selected tracker owns live status and execution history; otherwise
use compact continuation evidence alongside the plan. Scope changes route back
through the plan — hand-added tracker items never silently widen the
commitment.

## Reference Anchors

Use this local resource summary as grounding:

- `docs/resources/agile-manifesto-principles.md` grounds time-boxed
  commitment, sustainable cadence, and working evidence as the measure of
  progress.

## Focus

- Author a plan only when there is a commitment to coordinate — multiple
  owners, multiple workstreams, or a review with an audience. A solo
  iteration may run straight from the backlog selection and the tracker;
  the plan is coordination, not ceremony.
- Make the goal falsifiable: what is true at review, shown by what evidence.
- Commit exactly one Good, one Better, and one Best outcome per
  participating workstream. Good is the protected floor; Best drops first,
  then Better. More outcomes per workstream is confusing; fewer breaks the
  flex model — a workstream that cannot fill all three tiers sits this
  iteration out (not participating), never partially committed.
- Once committed, a Good outcome's ID, text, and acceptance evidence are
  immutable for the iteration. A Good that becomes impossible is an
  escalation recorded in the status report with its decider named — never an
  in-place edit, re-tier, or silent replacement.
- Give every outcome an owner, acceptance evidence, and at least one task.
  Outcome IDs are `<WS alias>-good|better|best`, unique within the plan;
  fully qualified as `<iteration-id>.<outcome-id>` when cited elsewhere.
- Reference workstreams by their roadmap-registry alias (`WS-<n>`) when a
  roadmap exists; a roadmap-less plan (single-workstream skip case) has one
  implicit workstream and uses no aliases — a second workstream means author
  the registry first. The plan never mints workstreams. Work items preserve
  the alias: as a label where the runtime supports labels (e.g. `ws:WS-1`,
  prefixed with the flow instance when several flows share one store),
  otherwise in the item's title or reference field.
- Keep the plan derivable: a runtime should be able to create work items
  from the task tables without inventing scope.
- Write tier labels in table cells as emoji plus a non-breaking space
  (U+00A0) plus the word — 🟢 Good, 🔵 Better, 🟣 Best — so the label never
  wraps apart in rendered tables. Plain words are fine in prose.
- Reference implementation plans where a committed outcome has one; do not
  duplicate their build slices here.

## Boundary Test

| If you are writing... | Put it in... |
|---|---|
| Workstream definitions and their `WS-<n>` aliases | Roadmap |
| Delivery sequence across iterations | Roadmap |
| Ranked improvement candidates with evidence | Improvement Backlog |
| One iteration's committed outcomes, owners, and dates | Iteration Plan |
| Design-derived build slices for one feature | Implementation Plan |
| Live task status, assignees, execution history | runtime work item or issue |
| How the iteration is going or went | Status Report |

## Completion Criteria

- The goal is falsifiable and the review date is set.
- Every participating workstream has exactly one Good, one Better, and one
  Best outcome, each with an owner and acceptance evidence.
- Every committed outcome maps to at least one task.
- Workstream aliases match the roadmap's registry (or the plan is
  roadmap-less and uses none).
- Trade rules make the drop order deterministic (Best, then Better; Good
  never).
- Every task has a stable ID, maps to a committed outcome, and names its
  optional work item, or explicitly uses direct plan execution.
Template
Show the template structure
---
ddx:
  id: iteration-plan
  authoring:
    home: repo
---

# Iteration Plan

| | |
|---|---|
| **Iteration** | [id — e.g. IT-05, Sprint 3] |
| **Dates** | [start → end] |
| **Review** | [date and format of the iteration review] |
| **Lead** | [who runs the iteration] |
| **Roadmap Slot** | [roadmap order/iteration this fulfills; omit when there is no roadmap] |

## Iteration Goal

> [One falsifiable sentence: what is true at review that is not true now,
> and what evidence shows it.]

## Committed Outcomes

Each participating workstream commits exactly one outcome per tier — one
Good, one Better, one Best. Good is the protected floor: the iteration fails
without it, and once committed its ID, text, and acceptance evidence are
immutable for the iteration — a Good that becomes impossible is recorded as
an escalation in the status report (decider named), never edited, re-tiered,
or silently replaced. Best drops first under pressure, then Better. More
outcomes per workstream is confusing; fewer breaks the flex model. A
workstream sitting out this iteration is listed as not participating, never
partially committed. (Participation is per-iteration; a workstream's
registry status — active/closed — lives in the roadmap and is a different
axis.) Outcome IDs are unique within this plan; the fully qualified form is
`<iteration-id>.<outcome-id>` (e.g. `IT-05.WS-1-good`).
Tier labels render as emoji plus a non-breaking space (U+00A0) so they never
wrap in table cells: 🟢 Good · 🔵 Better · 🟣 Best.

| ID | Workstream | Tier | Outcome | Owner | Governing Artifact | Acceptance Evidence |
|----|------------|------|---------|-------|--------------------|---------------------|
| [WS-1-good] | [WS-1] | 🟢 Good | [the outcome the iteration fails without] | [name] | [artifact ref] | [what proves it landed] |
| [WS-1-better] | [WS-1] | 🔵 Better | [real value, first traded after Best] | [name] | [artifact ref] | [what proves it landed] |
| [WS-1-best] | [WS-1] | 🟣 Best | [taken only if the iteration runs clean] | [name] | [artifact ref] | [what proves it landed] |

**Not participating this iteration**: [WS-n (reason), or None]

## Trade Rules

- Best drops first, then Better — per workstream. [Who] calls trades and
  records them in the status report.
- Good outcomes are never traded: a Good at risk is an escalation, not a
  trade.

## Tasks

The plan owns membership. Tasks can execute directly from the plan after user
authorization. The Work Item column is optional: use existing IDs or create
items only when tracking is explicitly selected. No item is required per task.
Every committed outcome maps to at least one task. Task IDs are stable —
never renumber mid-iteration — and unique within this plan (fully qualified:
`<iteration-id>.<task-id>`, e.g. `IT-05.1.1`). The Status column records
planning-time state (normally Backlog). When tracking is selected, the tracker
owns live status; otherwise keep recovery evidence alongside the plan.

Workstream aliases come from the roadmap's registry when one exists. A plan
authored without a roadmap (single-workstream skip case) has one implicit
workstream and omits aliases entirely; the moment a second workstream
appears, author the roadmap's registry first. Aliases are unique within one
flow instance. Work items preserve the alias: as a label where the runtime
supports labels (e.g. `ws:WS-1`, prefixed with the flow instance when
several flows share one store), otherwise in the item's title or reference
field.

### [WS-1 · workstream name]

| ID | Task | Outcome | Owner | Due | Status | Work Item |
|----|------|---------|-------|-----|--------|-----------|
| [1.1] | [task] | [WS-1-good] | [name] | [date] | [Backlog] | [optional item ID, or — for direct execution] |

## Risks

| Risk | Impact | Response |
|------|--------|----------|
| [risk] | [H/M/L] | [action] |
Innsigle seal: model-primary by HELIX

The signature covers the markdown source of this page, not these HTML bytes. This page quotes that seal; verify it against the source file.

Composition
model-primary
Issuer
HELIX helix
Signing key
ed25519:b0865d76d834a52c48506414d16f4e5a (build key)
Signed source
artifact-types/iterate/iteration-plan.md
Signed
2026-11-16T15:52:45Z
Content digest
sha256:793bfc9d…ee6cee48

This build key is endorsed by the human key for build signing; the signature is not a detector and not a truth guarantee.

Raw attestation JSON
{
  "payload": {
    "innsigle": "1",
    "type": "https://innsigle.dev/claim/colophon/v1",
    "issued_at": "2026-11-16T15:52:45Z",
    "issuer": {
      "id": "helix",
      "name": "HELIX",
      "key_id": "ed25519:b0865d76d834a52c48506414d16f4e5a",
      "key_url": "https://documentdrivendx.github.io/helix/.well-known/innsigle/keys.json"
    },
    "subjects": [
      {
        "uri": "https://documentdrivendx.github.io/helix/artifact-types/iterate/iteration-plan/",
        "digest": {
          "alg": "sha256",
          "value": "793bfc9d35919a0bf6e200a9c99ed5998b8af49cb83d3bb7e4a3ab83ee6cee48"
        }
      }
    ],
    "colophon": {
      "schema_version": "1",
      "composition": "model-primary",
      "ingredients": [
        {
          "kind": "model",
          "name": "Claude",
          "role": "draft"
        },
        {
          "kind": "tool",
          "name": "sloptimizer",
          "role": "rewrite"
        },
        {
          "kind": "human",
          "name": "operator",
          "role": "structure-edit"
        }
      ],
      "notes": null
    }
  },
  "payload_encoding": "json",
  "signatures": [
    {
      "key_id": "ed25519:b0865d76d834a52c48506414d16f4e5a",
      "alg": "ed25519",
      "sig": "AlQeHgOnCsAw-ymrxIZsygl5qwtwStPfhL8lhMdWrPvHbaJytmCQq5y_5ZEJXUX2iZtyXu3DIdj09ZcNL02zBw",
      "signed_at": "2026-11-16T15:52:45Z"
    }
  ]
}