Distill domain knowledge
Skill cat2000/domain-knowledge-ops/.cursor/skills/distill-domain-knowledge
Agent skills for Cursor: Confluence/Jira → domain briefs → story risk & INVEST splits and test case Design. Offline DEMO-1 — no Atlassian needed to try.
npx -y skills add cat2000/domain-knowledge-ops --skill distill-domain-knowledgeAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 24 days oldThe repository was created 24 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Advanced: Compose S3→S7 or redo Recognize/S2 when materialized/ already exists (no re-sync). Default resume after wiki prep is saying continue in the same thread — use this skill for partial steps or a fresh Compose session. Full pipeline: @generate-knowledge-from-wiki.
SKILL.md
2.5 KB, as published. Nobody here has run it
distill-domain-knowledge (Recognize or Compose — no Ingest)
Advanced resume. After wiki prep, the default is: human confirm → say continue (same thread as @generate-knowledge-from-wiki). Use this skill when you need Compose/Recognize without re-running Ingest, or a single-step / fresh-session resume.
Playbook: generate-knowledge-from-wiki/RUNBOOK.md (three stages; steps S1–S7). Chinese locale: RUNBOOK.zh-CN.md.
confirm / continue — same meaning as wiki skill. (zh-CN strings for these tokens: domain-knowledge/language/deliverable-locale-tokens.json.)
When to @
| Scenario | Do |
|---|---|
| Just finished wiki prep (same thread) | Wait for confirm, then say continue — do not require @distill |
| Checklist already confirm; new chat / no wiki thread | @distill-domain-knowledge <R> → Compose S3→S7 |
| Re-run Recognize only | @distill-domain-knowledge <R> + “S2” |
| One theme sub-step | @distill-domain-knowledge <R> + theme <slug> S3 / S4 / S5 / S6 / S7 |
Agent
- Open RUNBOOK and
domain-knowledge/distill-authoring-contract.md; run the requested range (S2:python3 scripts/distill/s2_recognize.py --root-id <R>). - S3 / S4 / S5 / S6: do not translate. S6 = source brief (
*-source-brief.md). S7 = locale / reader brief (*-domain-brief.mdfor en; zh-CN filename fromdeliverable-locale-tokens.json). - Keep industry axes: land evidence into strategy axes per
industry-axis-remount.md; do not promote source-tree layout into modules or hard-code tenant remount tables. - S2 contract: require
S2_DECISION_LEDGER.json+S2_REVIEW_DECISIONS.json; human entry isDOMAIN_MODULE_CHECKLIST.md. After S2:tagging_acceptance.py --root-id <R>before confirm. - After S3:
tagging_acceptance.py --root-id <R> --after-s3. After S7:--after-s7 --strict(zero-rule fake coverage = FAIL → revert confirm). domain_check distill: after S2 (coverage); after S7 (full).