Plans and executes bounded Lore memory recall with status checks, query expansion, context drill-down, freshness labels, cited evidence packets, and retrieval-failure recovery.
Use whenever an agent needs to answer from past Lore sessions, recall decisions, inspect last-session context, verify what happened before, or explain gaps without dumping transcripts.
Stronger alternatives
This repository is archived — consider an actively maintained alternative.
Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.
Claude CodeNot declared
CursorNot declared
CodexNot declared
GitHub CopilotNot declared
WindsurfNot declared
Gemini CLINot declared
ClineNot declared
OpenCodeNot declared
Repository health
Stars11
LicenseLICENSE
Default branchmain
Open issues14
Status
Archived
Package contents
Files included with this skill beyond the listing page.
skill mdSKILL.md3,353 B
docsSUMMARY.md380 B
History
First seen on skills.sh
First recorded snapshot · 2 installs
SKILL.md
Lore Recall
Use this workflow when past session memory matters. It sits above the low-level skills/lore CLI skill: use skills/lore for command mechanics and setup/indexing, use this skill for evidence planning, trust, freshness, and final packet shape.
Quick Start
Run lore status --json with the narrowest useful scope (--project, --source, --since, --until) before retrieval.
Build a recall plan: 2-4 query variants, likely filters, and why each query exists.
Search narrowly first: lore search "<terms>" --json, then --relevant when recency matters.
Drill down only from real ids returned by Lore: lore get <id> --full, lore context <id> --json, or lore session <session> --around <id> --json.
Emit a bounded evidence packet using references/evidence-packet.md.
If retrieval fails, report the gap and next queries tried. Do not invent memory.
Trust Rules
Lore transcripts are testimony, not truth. Current files, tests, runtime artifacts, and live system state beat stale transcript claims.
Never dump a whole session. Use search, one full message, or a bounded context/window.
Never invent ids. Every message/session id must come from Lore output.
Unknown source data stays null or "unknown". Do not infer branch, project, model, or freshness.
Real transcript excerpts are private. In committed examples/evals use synthetic ids and text only.
Freshness Labels
Use labels from references/freshness.md:
current: corroborated by current files/tests/runtime.
recent: transcript timestamp is plausibly relevant and no contradiction found.
stale: older memory may have drifted or current artifacts disagree.
unknown: missing timestamp or insufficient provenance.
Failure Recovery
If status says missingstore, emptystore, sourceabsent, possiblyunsynced, staleschema, or unreadablestore, stop and report the status plus recovery. If schemaVersion is greater than supportedSchemaVersion, say explicitly that read-only retrieval may work but write recovery such as lore sync can be refused until Lore is updated. If search returns zero hits, broaden once, try one synonym query, and report gaps instead of pretending.
References
references/evidence-packet.md for required output structure.
references/query-planning.md for query expansion and drill-down rules.