How do I install the Deep Recall skill?

Run npx skillmds@latest add modbender/deep-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 Deep Recall safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: executes scripts, makes network calls, reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Deep 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 Deep Recall free to use?

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

Who published Deep Recall?

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

Deep Recall by modbender — Agent Skill

Name: Deep Recall
Rating: 4.5 (4 reviews)
Author: modbender

File contents

DeepRecall — OpenClaw Skill

Description

Recursive memory recall for persistent AI agents using RLM (Recursive Language Models). Uses your existing OpenClaw LLM provider to recursively query memory files — no extra API keys needed.

When to Use

Agent needs to recall information from a large/growing memory store
Memory files exceed the context window (~128K tokens)
Agent needs to cross-reference across many daily log files
Deep reasoning about historical decisions, patterns, or events
Finding specific information across weeks/months of conversation history

Requirements

System Dependencies

Deno 2+ — runtime for fast-rlm
fast-rlm — the RLM engine. Clone it locally and set FAST_RLM_DIR=/path/to/fast-rlm
PyYAML (pip install pyyaml)

OpenClaw Configuration (Required)

A configured LLM provider in OpenClaw (Anthropic, OpenAI, Google, OpenRouter, Ollama, etc.)
The skill reads your OpenClaw config and credential files (~/.openclaw/) to auto-detect your provider and API key
No separate API key configuration needed — it reuses your existing OpenClaw setup

Environment Variables

FAST_RLM_DIR (required) — path to your cloned fast-rlm directory
OPENCLAW_WORKSPACE (optional) — override workspace path (defaults to ~/.openclaw/workspace)

What This Skill Accesses

DeepRecall reads local files and sends assembled context to your LLM provider:

What	Why	Where It Goes
`~/.openclaw/openclaw.json`	Find your LLM provider + model config	Stays local
`~/.openclaw/credentials/`	Read API keys (e.g. GitHub Copilot token)	Passed to fast-rlm process as env var
Workspace files (scope-dependent)	Build memory context for the query	Sent to your configured LLM provider via fast-rlm

Privacy Recommendations

Use identity or memory scope unless you need broader search
For maximum privacy, configure a local model provider (Ollama) so data never leaves your machine
Audit fast-rlm before running — it executes locally and receives your API key

Files

deep_recall.py — Main entry point (recall, recall_quick, recall_deep)
provider_bridge.py — Reads OpenClaw config to resolve API keys + models
model_pairs.py — Auto-selects cheap sub-agent model
memory_scanner.py — Discovers and indexes agent memory files
memory_indexer.py — Generates MEMORY_INDEX.md for efficient navigation
rlm_config_builder.py — Generates fast-rlm config

Quick Start

from deep_recall import recall
result = recall("What did we discuss last week about the project?")

Configuration

No additional API keys needed — DeepRecall reads your existing OpenClaw setup. Override RLM settings via config_overrides parameter.

---name: deep-recalldescription: Recursive memory recall for persistent AI agents using RLM (Recursive Language Models). Implements the Anamnesis Architecture — "The soul stays small, the mind scales forever."--- # DeepRecall — OpenClaw Skill ## Description Recursive memory recall for persistent AI agents using RLM (Recursive Language Models). Uses your existing OpenClaw LLM provider to recursively query memory files — no extra API keys needed. ## When to Use - Agent needs to recall information from a large/growing memory store - Memory files exceed the context window (~128K tokens) - Agent needs to cross-reference across many daily log files - Deep reasoning about historical decisions, patterns, or events - Finding specific information across weeks/months of conversation history ## Requirements ### System Dependencies - **[Deno](https://deno.com) 2+** — runtime for fast-rlm - **[fast-rlm](https://github.com/avbiswas/fast-rlm)** — the RLM engine. Clone it locally and set `FAST_RLM_DIR=/path/to/fast-rlm` - **PyYAML** (`pip install pyyaml`) ### OpenClaw Configuration (Required) - A configured LLM provider in OpenClaw (Anthropic, OpenAI, Google, OpenRouter, Ollama, etc.) - The skill **reads your OpenClaw config and credential files** (`~/.openclaw/`) to auto-detect your provider and API key - No separate API key configuration needed — it reuses your existing OpenClaw setup ### Environment Variables - `FAST_RLM_DIR` (required) — path to your cloned fast-rlm directory - `OPENCLAW_WORKSPACE` (optional) — override workspace path (defaults to `~/.openclaw/workspace`) ## What This Skill Accesses **DeepRecall reads local files and sends assembled context to your LLM provider:** | What | Why | Where It Goes | |------|-----|---------------| | `~/.openclaw/openclaw.json` | Find your LLM provider + model config | Stays local | | `~/.openclaw/credentials/` | Read API keys (e.g. GitHub Copilot token) | Passed to fast-rlm process as env var | | Workspace files (scope-dependent) | Build memory context for the query | Sent to your configured LLM provider via fast-rlm | ### Privacy Recommendations - Use **`identity` or `memory` scope** unless you need broader search - For maximum privacy, configure a **local model provider** (Ollama) so data never leaves your machine - **Audit [fast-rlm](https://github.com/avbiswas/fast-rlm)** before running — it executes locally and receives your API key ## Files - `deep_recall.py` — Main entry point (recall, recall_quick, recall_deep) - `provider_bridge.py` — Reads OpenClaw config to resolve API keys + models - `model_pairs.py` — Auto-selects cheap sub-agent model - `memory_scanner.py` — Discovers and indexes agent memory files - `memory_indexer.py` — Generates MEMORY_INDEX.md for efficient navigation - `rlm_config_builder.py` — Generates fast-rlm config ## Quick Start ```python from deep_recall import recall result = recall("What did we discuss last week about the project?") ``` ## Configuration No additional API keys needed — DeepRecall reads your existing OpenClaw setup. Override RLM settings via `config_overrides` parameter.

Frequently asked questions

Deep Recall