Story: Per-sprint capture workflow: individual files and compass integration
Table of Contents
This page documents a story in Sprint 18. It captures the goal, current status, acceptance criteria, and the tasks that compose it.
Goal
The single capture.org file per sprint has become unmanageable: annotations accumulate, implemented items are forgotten, and the file becomes noise rather than signal.
Replace it with product-level captures in the product backlog, where each capture is a separate file scaffolded by compass. New captures land in inbox/ (untriaged). At triage they move to next/ (imminent candidate), deferred/ (long-horizon idea), or discarded/ (explicitly rejected). The LLM calls compass capture --note "..." on the user's behalf; captures can then be promoted into stories or tasks. This makes triage cheap: promote the valuable, file the deferred, discard the stale.
Status
| Field | Value |
|---|---|
| State | DONE |
| Parent sprint | Sprint 18 |
| Now | Nothing. |
| Waiting on | — |
| Next | — |
| Last touched | 2026-05-29 |
Acceptance
compass capture --note "..."scaffolds a new capture indoc/agile/product_backlog/inbox/and prints the path.compass capture file <slug> next|deferred|discardedmoves a capture into the appropriate backlog bucket.compass capture promote <slug>prints instructions to promote a capture to a story or task.- Four buckets exist under
doc/agile/product_backlog/:inbox/,next/,deferred/, anddiscarded/, each with an index doc. - The product backlog uses
next/anddeferred/(renamed fromnear/andfar/). - The Product backlog knowledge doc documents all four buckets and the capture lifecycle.
- The Document types doc describes the capture type with the four-bucket structure.
- The commit recipe documents
[inbox],[next],[deferred],[discarded]as component brackets for backlog commits. - The Start work on a new story runbook links to the correct compass recipe and the product backlog doc.
- An org-roam capture template in
ores.lispwrites new captures to the inbox of the active environment. - A recipe documents how to write a freeform paragraph and have the LLM decompose it into well-structured captures in
inbox/. - All
near/farbucket references across the doc graph have been updated tonext/deferred. - Every entry in
sprint 18/capture.orghas been triaged into the appropriate product backlog bucket or discarded.
Tasks
| Task | State | Start | End | Description |
|---|---|---|---|---|
| Implement per-sprint capture workflow | DONE | 2026-05-27 | 2026-05-28 | Per-sprint capture workflow: individual files and compass integration |
| Update remaining near/far bucket references across all docs | DONE | 2026-05-27 | 2026-05-27 | Sweep all .org files for residual near/far bucket terminology and replace with inbox/next/deferred/discarded as appropriate. |
| Update document_types.org capture entry for inbox/next/deferred/discarded | DONE | 2026-05-27 | 2026-05-28 | Revise the capture document-type entry to describe the four-bucket structure (inbox, next, deferred, discarded) and per-file layout. |
| Write recipe: bulk capture from freeform text | DONE | 2026-05-27 | 2026-05-28 | Create a recipe for authoring a paragraph of freeform text and having the LLM decompose it into one or more well-structured captures filed in inbox/. |
| Add org-roam capture template in ores.lisp writing to inbox | DONE | 2026-05-27 | 2026-05-28 | Implement environment-aware org-capture template in ores.lisp that writes new captures directly to doc/agile/product_backlog/inbox/ of the current environment. |
| Add Emacs new-story and new-task commands with key bindings to ores.lisp | ABANDONED | 2026-05-27 | 2026-05-28 | Implement environment-aware new-story and new-task Emacs commands in ores.lisp, bound to keys. new-story prompts for title/description and scaffolds a story+task in the current sprint via compass. new-task lists current-sprint stories, lets user pick one, then scaffolds a task. Both commit with a well-formed message when the user confirms. |
| Apply bulk capture recipe to sprint 18 capture.org | DONE | 2026-05-28 | 2026-05-28 | Process sprint 18 capture.org using the bulk capture recipe: decompose entries into well-structured per-file captures filed in the product backlog inbox/next/deferred buckets. |
Decisions
Out of scope
- Migrating existing
capture.orgfiles to the new per-file layout (old files are left as-is; the new layout is forward-only). - Automatic background filing of captures without explicit user instruction.