メニュー
Queue schema for entity-mention signals that need later enrichment. Producer: OpenClaw Feishu signal-detector fork (external). Consumer: skills/kos-jarvis/enrich-sweep/run.ts. This file pins the on-disk format so both sides stay honest without shared code.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | pending-enrich |
| version | 1.0.0 |
| description | Queue schema for entity-mention signals that need later enrichment. Producer: OpenClaw Feishu signal-detector fork (external). Consumer: skills/kos-jarvis/enrich-sweep/run.ts. This file pins the on-disk format so both sides stay honest without shared code. |
| triggers | ["pending enrich","enrich queue"] |
| tools | [] |
| mutating | false |
Bridge between Phase 2 (Feishu ambient signal capture) and Phase 3
(batch enrich-sweep). See docs/JARVIS-NEXT-STEPS.md §4-§5 for the
phase context.
Path: ~/brain/agent/pending-enrich.jsonl
(override via PENDING_ENRICH_PATH env var)
Format: JSONL. One self-contained JSON object per line, newline-terminated. Never pretty-printed. Never rewritten — only appended to by producers, only rotated (move + truncate) by the consumer after a successful sweep.
{
"ts": "2026-04-17T10:32:14Z",
"name": "alice-example",
"kind": "person",
"context": "VC who invested in widget-co seed; mentioned by you on Feishu",
"source": "feishu:oc_abc123#msg_456",
"confidence": 0.7
}
ts — ISO 8601 UTC timestamp of the mentionname — entity name as spoken / written; do not slugify here,
dedup happens in enrich-sweepkind — one of person | company | concept | project. If unknown,
emit person and let enrich-sweep reclassifycontext — 40-300 chars of surrounding text. This becomes the first
line of the stub's State section if the entity gets promotedsource — opaque origin string. Convention:
feishu:<chat_id>#<message_id> for Feishu bridgemanual:<anything> for human-appended linesopenclaw:<skill-name> for other OpenClaw skillsconfidence — producer's self-reported 0-1 float. enrich-sweep
ignores <0.5 unless mention recursaliases — array of alternate strings observed alongside nameurl — canonical link if the mention came with one (tweet, article)O_APPEND
semantics (appendFileSync in Node) to avoid races.new Date().toISOString().mv ~/brain/agent/pending-enrich.jsonl \
~/brain/agent/pending-enrich.processing.<ts>.jsonl
.processing.* file (and may find another fresh one)..processing.<ts>.jsonl to
~/brain/agent/pending-enrich-archive/ after a clean run.{"ts":"2026-04-17T09:12:03Z","name":"Naval Ravikant","kind":"person","context":"Lucien discussed Naval's 'Specific Knowledge' framework","source":"feishu:oc_hub#msg_7781","confidence":0.9}
{"ts":"2026-04-17T09:15:44Z","name":"Tavily","kind":"company","context":"chose Tavily over Brave for enrich-sweep Tier 2 web search","source":"feishu:oc_hub#msg_7784"}
{"ts":"2026-04-17T11:02:11Z","name":"context-engineering","kind":"concept","context":"new concept page created in ideas/ during morning sync","source":"manual:lucien-terminal"}
POST /ingest on kos-compat-api (see
docs/FEISHU-SIGNAL-DETECTOR-SETUP.md).skills/kos-jarvis/enrich-sweep/SKILL.mddocs/FEISHU-SIGNAL-DETECTOR-SETUP.mdskills/signal-detector/SKILL.mdClassify vector-similar pages against each orphan page via Haiku 4.5 and add typed inbound references (link_type="related", relation encoded in context). Goal: chip away at the 1803/1952 orphan pile (v1 wiki flat import) across multiple sweep runs. Default is --dry-run; --apply does DB link inserts + markdown sentinel upsert on candidate pages that exist on disk, then commits ~/brain changes.
Merge WITHIN-SOURCE duplicate entity-variant pages (e.g. people/lucien + people/lucien-chen, companies/tp-link-system-inc + tp-link-systems-inc) into one canonical node: repoint links + facts + display-name aliases, register a slug redirect, soft-delete the alias pages. An LLM disambiguates each cluster (merge / ambiguous / distinct); ambiguous first-name-vs-multiple-full-names cases are quarantined, never merged. Default --dry (simulate + report, zero writes). Cross-source pages are NEVER touched — the two-axis design keeps them legitimately separate.
One-shot sweep that reads every page in the brain, extracts every person/company/concept/project mentioned via Haiku NER, dedupes, tiers by mention count + context, and creates stub pages for entities that lack one. Tier 2 candidates get Tavily web-search augmentation; Tier 3 stays brain-only (confidence: low). This is the primary G1 payoff skill from JARVIS-NEXT-STEPS.md §5.
Daily sweep: runs kos-lint + staleness detection + knowledge-gap analysis, then produces a MEMORY-format digest ready to be appended to OpenClaw's MEMORY.md by the digest-to-memory skill (Sun 22:00 weekly). Also writes a dashboard to ~/brain/agent/dashboards/knowledge-health-<date>.md.
Bridge an upstream gap: extract_atoms never stamps `frontmatter.concepts` on page-derived atoms, but synthesize_concepts consumes ONLY that field — so concept synthesis is permanently skipped ("no atoms with concept refs"). This script batch-labels atoms lacking the field with 1-3 kebab-case shared-vocabulary topic labels (sonnet) and stamps them, so upstream synthesize_concepts can tier + materialize concepts/<label> pages. Idempotent: only touches atoms WITHOUT `concepts`; safe to rerun after each extract_atoms drain. Retires when the upstream fix lands.
One-command bulk-ingest pipeline (D→I→K→W) for a single isolated source. Sequences the 7 fork stages — register → ingest → embed → enrich → graph → (per-entity synth, opt-in) → corpus-synth — over a folder of frontmatter'd .md files, each idempotent/incremental so re-runs only do new work. Turns a systematic corpus (product docs, FAQ, a research library) into a queryable knowledge asset: facts (RAG) + entities + graph + corpus-level viewpoints.