laqrumcode

Activate periodically (not per-session) to sample concepts from the graph and verify their claims against current state — code paths, recent sources, user-provided ground truth. Flag stale concepts for supersede-stale. Without this, knowledge rot poisons retrieval silently.

Activate when an insight worth keeping emerges mid-session — a surprising finding, a reusable pattern, a user correction, or a decision with rationale. Use this to write knowledge into the laqrumcode graph immediately rather than waiting for end-of-session daemon extraction (which is batch-only and can have hours of lag).

Activate when the user wants to distill lasting knowledge from ANY source — PDF, web article, pasted text, codebase file, transcript, markdown doc, or book chapter. Source-agnostic replacement for extract-pdf-gems. Same workflow core; source-type-aware quality rules.

Activate when the user asks to extract, mine, distill, or "pull gems" from a PDF, academic paper, book chapter, or research document into the laqrumcode memory graph. Also activate for phrases like "read this paper and remember it", "dense knowledge gems", "pull the insights", or when the user points at a PDF and asks to study it.

Activate when the user asks about state, history, prior work, codebase knowledge, or anything where injected memory should inform the response. Triggers include "what do you know about", "remember", "earlier", "last time", "we discussed", "in this codebase", "prior work", "previously", or any direct question about project/session state.

Activate when the user wants to know what they DO and DON'T know about a topic, before starting research or a project. Turns the graph from a reference library into an active planning tool by reporting coverage and explicit gaps.

Activate when user mentions "search memory", "recall", "remember", "core memory", "introspect", "memory status", "what did we do", "past sessions", "identity", "soul", or when you need to store/retrieve persistent knowledge across sessions.

Activate when the user wants to export laqrumcode for import into a non-SurrealDB system — Postgres + pgvector, Neo4j, OpenSearch, a custom store, or any system that ingests JSON. Triggers on "export laqrumcode to JSON", "dump laqrumcode for ingestion", "migrate laqrumcode off SurrealDB". Use this skill when the destination is not SurrealDB; for SurrealDB-to-SurrealDB use `laqrumcode-backup-native`.

Activate when the user asks to back up the laqrumcode database, export the graph, snapshot SurrealDB, or save a copy for restore into another SurrealDB instance. Triggers on "back up laqrumcode", "export the database", "snapshot memory", "dump laqrumcode to disk". Use this skill for the **lossless** path when the target is also SurrealDB.

Activate when the user wants to send the laqrumcode knowledge core (concepts, memories, skills, reflections, artifacts, soul) to another agent or system WITHOUT the transcript volume (turns, retrieval_outcomes, metrics). Triggers on "transfer knowledge to ilaqrum", "share laqrumcode brain", "extract just the concepts", "give my graph to another agent". For full snapshot use `laqrumcode-backup-native`; for non-SurrealDB targets use `laqrumcode-backup-jsonl`.

Selectively and REVERSIBLY forget stored memories/concepts (privacy / declutter) — soft-deactivate content matching a query or date so it stops surfacing in retrieval. Use when the user wants sensitive or unwanted data out of their laqrumcode graph.

Activate when the user asks about laqrumcode status, health, whether memory is working, before extracting knowledge from a new source, or when recall/write operations appear to be failing. Also activate for phrases like "is laqrumcode working", "check memory", "pipeline health", "anything broken".

End-to-end procedural skill for shipping a new laqrumcode version. Bumps all 6 version surfaces atomically, commits, tags, pushes, and verifies CI green with the correct exit-code pattern. Use BEFORE running `git push` on a release commit, not after.

Activate when the user wants to restore/import a laqrumcode JSON-Lines backup (produced by scripts/backup-jsonl.mjs) back into a SurrealDB laqrumcode graph, or merge one machine's laqrumcode export into another. Triggers on "restore laqrumcode from JSON", "import the jsonl backup", "merge laqrumcode graphs", "load my laqrumcode dump". Counterpart to laqrumcode-backup-jsonl.

Open the laqrumcode read-only web UI to explore the memory graph (dashboard, memory/concept browsers, graph explorer) in a browser. Use when the user asks to see/visualize/inspect what laqrumcode has stored.

Activate when raw recall results need interpretation — user asks a factual/historical question likely to have graph coverage, and you want clustered, contradiction-flagged, actionable output instead of a flat score-sorted list. Also activate when recall returns 5+ results and you need to decide which matter.

Activate when a recalled or injected concept is contradicted by current code, a newer source, or a user correction. Use this to demote stale knowledge in realtime rather than letting the batch daemon eventually catch it — stale concepts compete with fresh ones in recall and poison grounding.

Activate when the user wants a comparison, contrast, or synthesis across 2+ sources already in the graph. Produces meta-concepts that link back to the original source gems, earning the graph compound value from the cross-source edges.