agents/focus-area: auto-infer focus areas from task clustering #113

New Issue

alvis · 2026-05-05T09:26:02Z

alvis commented

2026-05-05 09:26:02 +00:00

Goal

Auto-infer per-user focus areas (named clusters of related work) from Todoist projects + labels, and write them into agent:focus-area preferences via the inference framework (#111).

Depends on

#97 (embedding-based task clustering — provides the nomic-embed-text clustering primitive this builds on).

Inferred parameters

Key	Shape	Default
`focusAreas`	`{ id, label, projectIds[], labelIds[], weight }[]`	empty
`primaryArea`	id of currently-most-active area	null

Method (sketch)

Embed task titles + project names with nomic-embed-text.
HDBSCAN or agglomerative cluster on recent (30d) tasks; weight by completion volume.
Generate human-readable labels for clusters via tip-generator LLM.
Surface top-N as suggested areas; user can confirm/rename/dismiss in admin (later).

Done when

Inference function plugged into the framework with TTL = 12h.
Cold-start uses no areas (agent emits a generic snippet).
FocusAreaAgent snippet reads primaryArea to bias the orchestrator prompt.
Tests cover: single-project user, many-project user, empty user.

## Goal Auto-infer per-user focus areas (named clusters of related work) from Todoist projects + labels, and write them into `agent:focus-area` preferences via the inference framework (#111). ## Depends on - #97 (embedding-based task clustering — provides the `nomic-embed-text` clustering primitive this builds on). ## Inferred parameters | Key | Shape | Default | |---|---|---| | `focusAreas` | `{ id, label, projectIds[], labelIds[], weight }[]` | empty | | `primaryArea` | id of currently-most-active area | null | ## Method (sketch) - Embed task titles + project names with `nomic-embed-text`. - HDBSCAN or agglomerative cluster on recent (30d) tasks; weight by completion volume. - Generate human-readable labels for clusters via `tip-generator` LLM. - Surface top-N as suggested areas; user can confirm/rename/dismiss in admin (later). ## Done when - Inference function plugged into the framework with TTL = 12h. - Cold-start uses no areas (agent emits a generic snippet). - `FocusAreaAgent` snippet reads `primaryArea` to bias the orchestrator prompt. - Tests cover: single-project user, many-project user, empty user.

alvis added this to the M2 — AI tips + multi-source signals milestone 2026-05-05 09:26:02 +00:00

alvis referenced this issue from a commit

2026-05-05 11:21:14 +00:00

feat(agents): per-agent inference — momentum, overdue-task, recent-patterns, focus-area (ADR-0014 step 7)

alvis closed this issue

2026-05-06 06:54:37 +00:00

alvis referenced this issue from a commit

2026-05-06 06:55:01 +00:00

feat(agents): semantic task clustering + focus-area inferred preferred_areas (#97, #113)

alvis referenced this issue from a commit

2026-05-06 06:56:22 +00:00

docs: update CLAUDE.md with session learnings (#97, #113)