How do I install the Episodic Recall skill?

Run npx skillmds@latest add proyecto26/episodic-recall in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Episodic Recall safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Episodic Recall?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Episodic Recall free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Episodic Recall?

proyecto26 (@proyecto26) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Episodic Recall by proyecto26 — Agent Skill

Name: Episodic Recall
Author: proyecto26

Episodic Recall

Queries episodic memory to find past experiences relevant to the current context. Use when any agent needs to recall past decisions, outcomes, or patterns from Jarvis's experience history. Currently a stub — full implementation pending Phase 2 Mem0 integration. Falls back to recall_similar_decisions tool.

proyecto26 27441a2 2.3 KB Updated Sep 17, 2026

File contents

Episodic Recall (Phase 2 Stub)

STATUS: STUB — This skill documents the Phase 2 episodic memory protocol. The full implementation requires recall_episodic tool backed by Mem0 + ChromaDB, which will be available in Phase 2. Until then, use the fallback below.

What This Will Do (Phase 2)

When fully implemented, this skill enables natural-language queries against Jarvis's episodic memory — all past journal entries, decisions, outcomes, and learnings indexed for hybrid semantic + keyword retrieval.

Phase 2 Protocol:

Agent formulates a natural language query about past experience
Query is sent to Mem0 (hybrid semantic + BM25 retrieval)
Top-k results are returned with metadata (date, section, cognitive_load_score)

Results are formatted as ActMem-style reasoning chains:

Context: <what was happening when this was stored>
Implication: <what this means for the current situation>
Confidence: <retrieval confidence score>

Agent uses these to inform current decision-making

Current Fallback

Until Phase 2 is available, use recall_similar_decisions tool instead:

recall_similar_decisions(situation="<describe the current situation>")

This searches the Judge's accumulated beliefs (flat JSON) for relevant entries. It's less powerful than the Phase 2 episodic memory but provides basic recall.

Limitations of fallback:

Only searches beliefs, not full journal history
No semantic search (exact/substring match only)
No confidence scoring
No temporal context (doesn't know WHEN beliefs were formed)

How to Use Now

Describe the current situation or question
Call recall_similar_decisions with that description
Review returned beliefs for relevance
Note: results will be much richer after Phase 2 migration

Phase 2 Prerequisites

mem0ai package installed
chromadb running (embedded mode)
Journal entries indexed via EpisodicMemory.index_journal_entry()
recall_episodic tool registered in agent tools

---name: episodic-recalldescription: Queries episodic memory to find past experiences relevant to the current context. Use when any agent needs to recall past decisions, outcomes, or patterns from Jarvis's experience history. Currently a stub — full implementation pending Phase 2 Mem0 integration. Falls back to recall_similar_decisions tool.---# Episodic Recall (Phase 2 Stub)> **STATUS: STUB** — This skill documents the Phase 2 episodic memory protocol.> The full implementation requires `recall_episodic` tool backed by Mem0 + ChromaDB,> which will be available in Phase 2. Until then, use the fallback below.## What This Will Do (Phase 2)When fully implemented, this skill enables natural-language queries against Jarvis'sepisodic memory — all past journal entries, decisions, outcomes, and learningsindexed for hybrid semantic + keyword retrieval.**Phase 2 Protocol:**1. Agent formulates a natural language query about past experience2. Query is sent to Mem0 (hybrid semantic + BM25 retrieval)3. Top-k results are returned with metadata (date, section, cognitive_load_score)4. Results are formatted as ActMem-style reasoning chains: ``` Context: <what was happening when this was stored> Implication: <what this means for the current situation> Confidence: <retrieval confidence score> ```5. Agent uses these to inform current decision-making## Current FallbackUntil Phase 2 is available, use `recall_similar_decisions` tool instead:```recall_similar_decisions(situation="<describe the current situation>")```This searches the Judge's accumulated beliefs (flat JSON) for relevant entries.It's less powerful than the Phase 2 episodic memory but provides basic recall.**Limitations of fallback:**- Only searches beliefs, not full journal history- No semantic search (exact/substring match only)- No confidence scoring- No temporal context (doesn't know WHEN beliefs were formed)## How to Use Now1. Describe the current situation or question2. Call `recall_similar_decisions` with that description3. Review returned beliefs for relevance4. Note: results will be much richer after Phase 2 migration## Phase 2 Prerequisites- `mem0ai` package installed- `chromadb` running (embedded mode)- Journal entries indexed via `EpisodicMemory.index_journal_entry()`- `recall_episodic` tool registered in agent tools

Frequently asked questions