wiki-session-memory
Skill0 starsChanged 4 months agoMITView on GitHub
Per-chat session memory in raw/memory/ — save, recall, list, prune. Use when the user wants continuity across Claude sessions or to audit past work.
- Installs packages
In this repository36 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- wiki-extract-annasSkill
- wiki-extract-crunchbaseSkill
- wiki-extract-ebookSkill
- wiki-extract-ecommerceSkill
- wiki-extract-githubSkill
- wiki-extract-linkedinSkill
- wiki-extract-newsletterSkill
- wiki-extract-patentsSkill
- wiki-extract-paywallSkill
- wiki-extract-podcastSkill
- wiki-extractSkill
- wiki-extract-wikipediaSkill
- wiki-extract-youtubeSkill
- wiki-fetchSkill
- wiki-ingestSkill
- wiki-learnSkill
- wiki-lintSkill
- wiki-maintainerSkill
- wiki-onboardSkill
- wiki-pipelineSkill
- wiki-querySkill
- wiki-raw-prepareSkill
- wiki-research-academicSkill
- wiki-research-deepSkill
- wiki-research-feedsSkill
- wiki-research-loopSkill
- wiki-research-newsSkill
- wiki-researchSkill
- wiki-research-socialSkill
- wiki-research-webSkill
- wiki-retroSkill
- wiki-setupSkill
- wiki-statusSkill
- wiki-upgradeSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Read llm-wiki/config.json.
- Very specific: line 16
Prefer shell commands; use MCP tools only when the agent has no shell (skills/references/mcp-and-kg.md § "CLI > MCP when local").
- Very specific: line 39
When memory.max_sessions is a positive integer, memory save and memory log run an automatic prune after each write: the oldest session …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Wiki session memory
Stores one markdown file per chat session: raw/memory/<session-id>.md. Optional feature: set memory.enabled to true in llm-wiki/config.json.
Pre-flight
- Read
llm-wiki/config.json. Ifmemory.enabledis false, say so and skip writes (read-only list/show may still work for existing files).
CLI first
Prefer shell commands; use MCP tools only when the agent has no shell (skills/references/mcp-and-kg.md § "CLI > MCP when local").
llm-wiki memory save --current --summary "What we did" --tags "research,ingest"
llm-wiki memory recall "topic" --tag research --limit 5
llm-wiki memory list --json
llm-wiki memory show --current
llm-wiki memory prune --keep 30--current (or -c) reads llm-wiki/.current-session (updated by Claude Code hooks).
Relationship to wiki-learn
llm-wiki/.agent-memory.md— distilled cross-session patterns (wiki-learn).raw/memory/*.md— per-session raw records; promote recurring patterns into.agent-memory.mdwhen appropriate.
Recall quality
memory recall uses the vault search backend (mcp.search_backend in config.json — fts5, grep, chromadb, or hybrid). Results match how wiki_search / CLI search would rank the same text: keyword-heavy notes work well with fts5; broader or conceptual queries improve with chromadb / hybrid (optional pip install chromadb). See skills/references/mcp-and-kg.md § Search backends.
memory.max_sessions
When memory.max_sessions is a positive integer, memory save and memory log run an automatic prune after each write: the oldest session files (by file modification time) are deleted until the count is ≤ the cap. Set 0 for unlimited files (manual memory prune only).
Security and privacy
Session files may be git-tracked if the user commits raw/. Do not store API keys, tokens, or other secrets in raw/memory/*.md — keep credentials in gitignored config (e.g. .claude/settings.local.json) or env vars per docs/ENV.md.
Done looks like
- User knows whether memory is enabled and where files live (
raw/memory/). - For writes:
memory saveran with--currentor an explicit session id.
Related skills
- wiki-learn — cross-session distilled memory
- wiki-query — can
memory recallwhen wiki lacks an answer - wiki-research — recall past sessions on the same topic in pre-flight
Smoke check
- CLI:
llm-wiki memory listfrom vault root. - Prompt: slash command
/llm-wiki:memoryor opencommands/memory.md. - Config:
memory.enabled,memory.dir,memory.max_sessionsinllm-wiki/config.json.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

