Skip to content
Skills Library
AI & agents · Planning

classify

/classify

Propose work items from recent conversation (flag-gated classifier). Use when the Stop hook requests classification or when asked to sweep a conversation for untracked work. Propose-only: writes .work/suggestions.jsonl, never the event log.

Licence
MIT
Last repository change
22.09.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 22.09.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Propose work items from recent conversation (flag-gated classifier). Use when the Stop hook requests classification or when asked to sweep a conversation for untracked work. Propose-only: writes .work/suggestions.jsonl, never the event log.

Original author description

Propose work items from recent conversation (flag-gated classifier). Use when the Stop hook requests classification or when asked to sweep a conversation for untracked work. Propose-only: writes .work/suggestions.jsonl, never the event log.

How it works

  1. 1. Spawn ONE background subagent
  2. 2. Dispose next turn (main loop)
Original SKILL.md headings · 50 lines, 305 words
SKILL.md preview
---
name: classify
metadata:
  version: "0.24.12"
description: "Propose work items from recent conversation (flag-gated classifier). Use when the Stop hook requests classification or when asked to sweep a conversation for untracked work. Propose-only: writes .work/suggestions.jsonl, never the event log."
---

# Classify — propose work items from conversation

Flag-gated (spec §6): meaningful only when `classifier.enabled: true` in
`.work/config.yml`. Propose, don't dispose — suggestions go to
`.work/suggestions.jsonl` (gitignored, per-clone, append-only); nothing enters
the event log until the main loop or the human promotes.

## 1. Spawn ONE background subagent

Spawn a single background subagent (never more than one) with this prompt:

> Analyze the recent exchange for trackable work that has no work item yet.
> For each plausible item, append ONE JSON line to `.work/suggestions.jsonl`:
>
>     {"suggestion_id": "<ULID>", "source_span": "<turn or transcript ref>",
>      "proposed": {"level": "task", "kind": "feature", "parent": null,
>                   "milestone": null, "title": "..."},
>      "confidence": 0.82,
>      "open_questions": [], "dedupe_against": ["<item ULIDs checked>"]}
>
> - Mint each `suggestion_id` fresh:
>   `python3 -c "import sys; sys.path.insert(0,'bin'); import ulid; print(ulid.new())"`
> - Dedupe FIRST: check the current `bin/worklog fold` titles AND prior
>   unconsumed suggestions already in `.work/suggestions.jsonl`. Never
>   re-propose an item that exists or was already suggested; record what you
>   checked in `dedupe_against`.
> - Read `classifier.min_confidence` from `.work/config.yml` (default 0.7).
>   If confidence < min_confidence, `proposed.kind` MUST be `"triage"` and the
>   doubt goes in `open_questions` — never a confident guess on a load-bearing
>   field.
> - NEVER append to `.work/todo.jsonl`, `.work/done.jsonl`, or any event log.
>   Suggestions only.
> - NEVER ask the user anything — you cannot block. Write suggestions and exit.

## 2. Dispose next turn (main loop)

Next turn, read `.work/suggestions.jsonl`, skipping records already marked
consumed:

- High-confidence suggestions: offer them for promotion —
  `bin/worklog promote <suggestion-id>` creates the real item.
- Low-confidence (`kind:triage`) suggestions: surface their `open_questions`
  to the user as questions, not items.
- Rejected suggestions: mark consumed by appending
  `{"consumed": true, "suggestion_id": "..."}` to `.work/suggestions.jsonl`.

The subagent gathers and proposes; the main loop, which can talk to the human,
disposes.

Versions

RepositoryScoreLicence
This page · spillwavesolutions/wiki_ticket_sdd90MIT
synaptiai/agent-capability-standard100Apache-2.0

Identical content appears in 1 other GitHub folders. We show one copy with the highest score.

Database snapshot: 13.11.2026 · package version c2d19e1b37ca (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90apply-chat-suggestions
    Address suggestions in a past conversation's insight by performing each suggestion's actionable fix. Use when the user asks to implement, apply, address, or act on the suggestions from a conversation insight. Requires a conversation ID. Do NOT use this to generate a new insight.
  • 100conversation-analysis
    Classify an AI assistant's conversations by use case, resolution and sentiment, PII masked. Use to build or check the taxonomy, pick the classifier, see what people use it for, or act on the results.
  • 100jsonl-core
    This skill should be used when the user asks to "analyze conversation history", "parse JSONL files", "read past sessions", "search conversation logs", "find what happened in a session", or needs to work with Claude Code's .jsonl conversation format. It provides the canonical…
  • 95conversation
    Use when interacting with users — asking clarifying or exploratory questions, presenting suggestions or options, pausing for confirmation, or before taking an action with effects outside the repository.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.