Agent Plugins Marketplace
← All plugins

memento

v0.19.12

Multi-layer cache memory base. L3 (sources) -> L2 (wiki) -> L1 (AGENTS.md) compilation pipeline with a hardened deterministic eval gate (source evidence proves provenance; compiled answer surfaces must contain required atoms; ungated runs are explicit), health checks, passive session capture (save), active LLM-driven interview (ama), and lightweight follow-up queue review (followups). Knowledge base only; task storage is an explicit root-level opt-in.

CodexClaude Code6 Skills

By Rich HaaseLicense: MIT0 GitHub starsUpdated 1 hour ago

Directory evidence

Runtimes
Codex and Claude Code
Parsed components
6 skill or MCP entries
Source updated
Sep 26, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install memento for Codex and Claude Code

Installs for the current user
codex plugin marketplace add richhaase/toolshed
codex plugin marketplace upgrade toolshed
codex plugin add memento@toolshed

Paste and run these commands in a terminal with Codex. They add and refresh the toolshed catalog, then install this plugin.

Compatibility: the page URL and API slug “memento” remain stable.

  • Codex: memento@agent-plugin-marketplace → memento@toolshed

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/richhaase/toolshed

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/memento/.

Plugin files

plugins/memento/
├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── skills/ama/SKILL.md
├── skills/compile/SKILL.md
├── skills/followups/SKILL.md
├── skills/health-check/SKILL.md
├── skills/memento-config/SKILL.md
└── skills/save/SKILL.md

Included Skills6

amaskills/ama/SKILL.md

Read the wiki and interview the user to fill gaps, resolve contradictions, and capture context the LLM is curious about. Use when the user says "ama", "ask me anything", "what do you want to know", "interview me", "fill in gaps", "tell me what's missing", or otherwise wants the agent to actively pull knowledge into the Memento. Captures answers as a session source so /compile can fold them into the wiki. Not `save`, which extracts what already happened in a session, and not `compile`, which synthesizes existing sources into the wiki.

compileskills/compile/SKILL.md

Use when the user says "compile", "update the wiki", "build wiki", "compile wiki", or wants to refresh the memory base. Reads sources/**/*.md and synthesizes into wiki/ organized by topic/entity, then refreshes the AGENTS.md hot set. First run builds full wiki; subsequent runs do incremental updates. Not for browsing open items (see `followups`), closing a session (see `save`), or filling gaps (see `ama`).

followupsskills/followups/SKILL.md

Inventory, view, or triage the Memento follow-up queue. Use when the user names Memento follow-ups, asks to list/review this Memento's open items, says "show me followup X", or wants to walk the queue one item at a time (keep, dismiss, answer, note, or file-and-dismiss). List mode is expired-first. Do not trigger on generic task, issue, PR, email, or "what's open" requests without Memento context.

health-checkskills/health-check/SKILL.md

Audit a Memento for cache drift, stale projections, broken evidence paths, privacy-boundary risks, compile metadata drift, open queue visibility, or golden-query eval readiness. Use when the user asks for a health check, doctor, audit, staleness check, provenance check, privacy lint, L1/L2/L3 integrity check, or whether a Memento is current. Read-only by default and focused on public surfaces. Not for auditing an Agent Skill, plugin, or marketplace against the skills spec — that is `actuary:skill-audit`.

memento-configskills/memento-config/SKILL.md

Set up a new Memento or reconfigure an existing one. Use when the user asks to create/scaffold a Memento knowledge base, initialize its wiki and canonical AGENTS.md context, repair missing structure, or change entity types, profile, nicknames, or labels. Detects new versus existing state and applies targeted, idempotent updates. This is the Memento entry point, not the workflow for compiling, capturing a session, triaging follow-ups, or running diagnostics.

saveskills/save/SKILL.md

Save a substantive session into the Memento by extracting durable context from the visible conversation and returned agent results. Use when the user explicitly asks to save, log, or capture this session in the Memento, including "anything worth capturing?" Default mode writes immediately; `ask` reviews the plan first. Do not trigger from a generic thanks, sign-off, wrap-up, or close-out without an explicit persistence request.

Plugin manifests2

plugins/memento/.codex-plugin/plugin.json
{
  "name": "memento",
  "version": "0.19.12",
  "description": "Multi-layer cache memory base. L3 (sources) -> L2 (wiki) -> L1 (AGENTS.md) compilation pipeline with a hardened deterministic eval gate (source evidence proves provenance; compiled answer surfaces must contain required atoms; ungated runs are explicit), health checks, passive session capture (save), active LLM-driven interview (ama), and lightweight follow-up queue review (followups). Knowledge base only; task storage is an explicit root-level opt-in.",
  "author": {
    "name": "Rich Haase",
    "url": "https://github.com/richhaase"
  },
  "repository": "https://github.com/richhaase/toolshed",
  "license": "MIT",
  "keywords": [
    "memory",
    "memento",
    "agent-memory",
    "knowledge-preservation",
    "wiki",
    "notes",
    "followups",
    "compile",
    "interview",
    "personal-knowledge-management"
  ],
  "skills": "./skills/",
  "interface": {
    "composerIcon": "./assets/icon.png",
    "logo": "./assets/icon.png",
    "displayName": "Memento",
    "shortDescription": "Personal memory base workflows",
    "longDescription": "Multi-layer local memory base. Six skills cover the full loop: memento-config (setup + ongoing config with exact-path commit safety), health-check (read-only doctor + golden-query scorer), save (session capture), ama (active interview), compile (sources -> wiki -> AGENTS.md, deterministic eval plus snapshot-backed rollback that preserves pre-existing work), and followups (small queue review). Task storage is disabled by default and may be explicitly enabled by a root policy. Supports a configurable Memento root for global installs and treats ingested content as untrusted evidence.",
    "developerName": "Rich Haase",
    "category": "Productivity",
    "capabilities": [
      "Read",
      "Write"
    ],
    "websiteURL": "https://github.com/richhaase/toolshed",
    "defaultPrompt": [
      "Set up a memory base in this repo.",
      "Run a health check on my memory base.",
      "Compile my sources into the wiki.",
      "Save this session and capture useful context.",
      "Ask me anything to fill in gaps in my wiki."
    ]
  }
}
plugins/memento/.claude-plugin/plugin.json
{
  "name": "memento",
  "version": "0.19.12",
  "description": "Multi-layer cache memory base. L3 (sources) -> L2 (wiki) -> L1 (AGENTS.md) compilation pipeline with a hardened deterministic eval gate (source evidence proves provenance; compiled answer surfaces must contain required atoms; ungated runs are explicit), health checks, passive session capture (save), active LLM-driven interview (ama), and lightweight follow-up queue review (followups). Knowledge base only; task storage is an explicit root-level opt-in.",
  "author": {
    "name": "Rich Haase",
    "url": "https://github.com/richhaase"
  },
  "repository": "https://github.com/richhaase/toolshed",
  "license": "MIT",
  "keywords": [
    "memory",
    "memento",
    "agent-memory",
    "knowledge-preservation",
    "wiki",
    "notes",
    "followups",
    "compile",
    "interview",
    "personal-knowledge-management"
  ],
  "skills": "./skills/"
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[memento on Agent Plugins Marketplace](https://pluginsmp.com/plugins/memento)