Dashboard: Artifact organization by type and phase with relevance ordering #351

New issue

Open

opened 2026-04-15 08:44:09 +00:00 by mprihoda · 1 comment

mprihoda commented

2026-04-15 08:44:09 +00:00

(Migrated from github.com)

Layer: 2 — Navigation & Document Experience
Parent: #343

Problem

Artifacts are listed in a flat, unstructured way. When a worktree has many artifacts across multiple phases (analysis, review packets, release notes, diagrams), finding the right document requires scanning the full list. Artifacts also accumulate duplicates without relevance ordering (#200).

Solution

Organize artifacts by type (analysis, phase context, review packet, release notes, diagrams) and/or by phase number
Relevance ordering — most recent/relevant artifacts at the top, stale ones deprioritized
Deduplicate — when artifacts are regenerated, show the latest version prominently

Incorporates #200 (artifacts accumulate with duplicates and lack relevance ordering)

**Layer:** 2 — Navigation & Document Experience **Parent:** #343 ## Problem Artifacts are listed in a flat, unstructured way. When a worktree has many artifacts across multiple phases (analysis, review packets, release notes, diagrams), finding the right document requires scanning the full list. Artifacts also accumulate duplicates without relevance ordering (#200). ## Solution 1. **Organize artifacts by type** (analysis, phase context, review packet, release notes, diagrams) and/or by phase number 2. **Relevance ordering** — most recent/relevant artifacts at the top, stale ones deprioritized 3. **Deduplicate** — when artifacts are regenerated, show the latest version prominently ## Related - Incorporates #200 (artifacts accumulate with duplicates and lack relevance ordering)

claude commented

2026-10-16 05:55:09 +00:00

Member

Intake: needs-input
Workflow: -
Reason: the goal is not specified — the deduplication part is done (IW-715, 771e37ba), but the issue does not choose how to group, what "relevant" means, or which view changes
Analysis: - (no branch, no worktree)

1. What must group the artifacts, and where does the group of each artifact come from?

Each artifact on the card has only a label and a path (ReviewState.scala#L15-L18). The schema permits an optional category field (review-state.schema.json), but the dashboard does not read it. The issue says:

Organize artifacts by type (analysis, phase context, review packet, release notes, diagrams) and/or by phase number

a) Group by phase, then by type in each phase. Get both from the file name (phase-02-context.md, review-packet-phase-01.md). This needs no schema change and no workflow change. A file name that does not match a rule goes into an "Other" group.
b) The same groups, but each workflow writes the type into category. The groups are exact. This changes the plugin skills and old cards show no groups.
c) Group by type only.

Recommended: a): it works on the cards that exist now, and it changes the least.

2. What does "most relevant first" mean?

The dashboard shows artifacts in the order of the array. A replaced artifact keeps its position, so the array order is not the order of recency. The card records the current phase. Issue #200 (open) also proposes a separate attention_artifacts list that the workflow sets.

a) The current phase first, then the earlier phases newest first, then the documents that have no phase. This uses the phase field that the card already has.
b) Sort by file modification time. This needs a file-system read for each artifact on each render.
c) The workflow sets an attention list (#200 part 2). This needs a schema change and plugin skill changes.

Recommended: a): it uses data that is already on the card and it gives #200 part 2 no work to do.

3. Which view must show the grouped list?

The worktree card and the detail page use the same flat list (WorktreeCardRenderer.scala#L934-L949). The effort map puts this work after the sidebar of #350 (open). The #350 sidebar also lists "artifacts organized by type/phase".

a) Change the list that the card and the detail page use now. It ships now. Later, #350 uses the same grouping.
b) Wait for #350 and group only in its sidebar. This blocks this issue on #350.
c) The detail page only. The card keeps a short list of the current phase.

Recommended: a): one grouping function serves all views, and nothing waits.

Answer in a comment on this issue, per number, for example "1: a", or in your own words; intake reads it on its next iteration.

Intake: needs-input Workflow: - Reason: the goal is not specified — the deduplication part is done (IW-715, 771e37ba), but the issue does not choose how to group, what "relevant" means, or which view changes Analysis: - (no branch, no worktree) ### 1. What must group the artifacts, and where does the group of each artifact come from? Each artifact on the card has only a label and a path ([ReviewState.scala#L15-L18](https://code.iterative.works/iterative-works/iw-cli/src/branch/main/core/model/ReviewState.scala#L15-L18)). The schema permits an optional `category` field ([review-state.schema.json](https://code.iterative.works/iterative-works/iw-cli/src/branch/main/schemas/review-state.schema.json)), but the dashboard does not read it. The issue says: > Organize artifacts by type (analysis, phase context, review packet, release notes, diagrams) and/or by phase number - **a) Group by phase, then by type in each phase. Get both from the file name** (`phase-02-context.md`, `review-packet-phase-01.md`). This needs no schema change and no workflow change. A file name that does not match a rule goes into an "Other" group. - **b) The same groups, but each workflow writes the type into `category`.** The groups are exact. This changes the plugin skills and old cards show no groups. - **c) Group by type only.** **Recommended: a)**: it works on the cards that exist now, and it changes the least. ### 2. What does "most relevant first" mean? The dashboard shows artifacts in the order of the array. A replaced artifact keeps its position, so the array order is not the order of recency. The card records the current `phase`. Issue #200 (open) also proposes a separate `attention_artifacts` list that the workflow sets. - **a) The current phase first, then the earlier phases newest first, then the documents that have no phase.** This uses the `phase` field that the card already has. - **b) Sort by file modification time.** This needs a file-system read for each artifact on each render. - **c) The workflow sets an attention list (#200 part 2).** This needs a schema change and plugin skill changes. **Recommended: a)**: it uses data that is already on the card and it gives #200 part 2 no work to do. ### 3. Which view must show the grouped list? The worktree card and the detail page use the same flat list ([WorktreeCardRenderer.scala#L934-L949](https://code.iterative.works/iterative-works/iw-cli/src/branch/main/dashboard/jvm/src/presentation/views/WorktreeCardRenderer.scala#L934-L949)). The effort map puts this work after the sidebar of #350 (open). The #350 sidebar also lists "artifacts organized by type/phase". - **a) Change the list that the card and the detail page use now.** It ships now. Later, #350 uses the same grouping. - **b) Wait for #350 and group only in its sidebar.** This blocks this issue on #350. - **c) The detail page only. The card keeps a short list of the current phase.** **Recommended: a)**: one grouping function serves all views, and nothing waits. Answer in a comment on this issue, per number, for example "1: a", or in your own words; intake reads it on its next iteration.

claude added the

needs-input

label

2026-10-16 05:55:09 +00:00

claude referenced this issue

2026-10-16 07:25:14 +00:00

review-state: Artifacts accumulate with duplicates and lack relevance ordering #200