Skip to content
Back to skills

Rlm Init

ASecurity

Interactive RLM cache initialization for setting up a project semantic cache or adding a profile.

  • 7 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 2, 2026
ai-agentspythonbashnodedatabase

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned October 3, 2026

npx -y skills add richfrem/agent-plugins-skills --skill rlm-init --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Rlm Init?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Rlm Init
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/richfrem-rlm-init-agent-plugins-skills/badge)](https://www.skillsdirectory.com/skills/richfrem-rlm-init-agent-plugins-skills)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: rlm-init
plugin: agent-memory
description: Interactive RLM cache initialization for setting up a project semantic cache or adding a profile.
allowed-tools: Bash, Read, Write
---

# RLM Cache Initialization (`rlm-init`)

Initializes a new Recursive Language Model (RLM) semantic cache for zero-dependency high-speed memory.

## Contents

- [Critical Constraints](#critical-constraints)
- [Quick start](#quick-start)
- [Setup Modes](#setup-modes)
- [Workflow](#workflow)
- [Verification](#verification)
- [References](#references)

## Critical Constraints

1. **Zero-Dependency**: Uses standard library Python 3.8+ only; no external databases or daemons required.
2. **Profile Storage**: Profiles default to `.agent/learning/rlm_profiles.json`.
3. **Scripted Writes**: Never create cache markdown files manually; use `scripts/inject_summary.py`.

## Quick start

Check existing coverage or initialize a profile:

```bash
python3 scripts/inventory.py --profile project
```

## Setup Modes

- **Mode A (Standalone)**: RLM only. Fast keyword lookup across dense file summaries.
- **Mode B (RLM + vector-db)**: RLM keyword pre-filter into vector semantic search.
- **Mode C (RLM + obsidian-wiki-engine)**: RLM as wiki concept distiller.
- **Mode D (Super-RAG)**: Full stack combining RLM, vector store, and wiki concept nodes.

## Workflow

1. **Requirements Discovery**: Select target directories, file types, and cache destination.
2. **Profile Configuration**: Define profile entry in `.agent/learning/rlm_profiles.json`.
3. **Manifest Creation**: Define include/exclude patterns in `<name>_manifest.json`.
4. **Audit Gaps**: Run `python3 scripts/inventory.py --profile <name>` to list uncached files.
5. **Initial Distillation**: Deep read and inject summaries via `python3 scripts/inject_summary.py`.
6. **Report Readiness**: Confirm initialized profile paths and ready state to the user.

## Verification

Confirm 100% cache coverage with the inventory script:

```bash
python3 scripts/inventory.py --profile project
```

## References

- [acceptance-criteria.md](references/acceptance-criteria.md) — Acceptance criteria for RLM cache bootstrapping.
- [rlm-bootstrap-guide.md](references/rlm-bootstrap-guide.md) — Comprehensive bootstrap manual and mode details.

Files in this skill

  • SKILL.md7.2 KB
  • acceptance-criteria.md440 B
  • assets/resources/distiller_manifest.json52 B
  • assets/resources/manifest-index.json48 B
  • assets/resources/prompts/rlm/rlm_summarize_general.md71 B
  • assets/resources/prompts/rlm/rlm_summarize_tool.md68 B
  • assets/resources/rlm_manifest.json46 B
  • assets/rlm_profiles.json53 B
  • assets/rlm_summary_cache_manifest.json67 B
  • assets/rlm_tools_manifest.json59 B
  • evals/evals.json693 B
  • evals/results.tsv172 B
  • references/acceptance-criteria.md42 B
  • references/cheapest_models.json40 B
  • references/cheapest_models.md38 B
  • requirements.txt22 B
  • scripts/cleanup_cache.py33 B
  • scripts/distiller.py29 B
  • scripts/inject_summary.py34 B
  • scripts/inventory.py29 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…