Skip to content
Back to skills

Docs

ASecurity

This is the companion skill doc for the `skills/memory-layer/SKILL.md` extension registration. For the authoritative and complete skill reference, see [`../skills/memory-layer/SKILL.md`](../skills/memory-layer/SKILL.md).

  • 54 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 28, 2026
ai-agentspythonrustgosqlrailsapibackend

Works with

  • claude code
  • cli
  • api
  • mcp

Security analysis

A100/100

Scanned September 28, 2026

npx -y skills add GeneGulanesJr/LaPis --skill docs --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Docs?

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

Security grade badge for Docs
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/genegulanesjr-docs/badge)](https://www.skillsdirectory.com/skills/genegulanesjr-docs)

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

SKILL.md
# Memory Layer Skill

This is the companion skill doc for the `skills/memory-layer/SKILL.md` extension registration. For the authoritative and complete skill reference, see [`../skills/memory-layer/SKILL.md`](../skills/memory-layer/SKILL.md).

## Overview

The memory layer provides persistent memory capabilities for the Pi coding agent. It runs as a Pi extension with a local SQLite backend, providing observation CRUD, code indexing, code analysis, doc indexing, trust scoring, and session lifecycle management — all without cloud dependencies or API keys.

The same backend also powers a [Claude Code CLI](https://code.claude.com/docs/en/overview) integration (`lapis claude-code install`) that wires MCP tools plus lifecycle hooks for context injection, guardrails, passive capture, and session tracking. See [`CLAUDE_CODE.md`](CLAUDE_CODE.md).

## Key Capabilities

- **Declarative memory** — save, search, update, and delete observations (decisions, bugfixes, patterns, discoveries).
- **Code indexing** — web-tree-sitter (WASM) parses JS/TS/TSX/Go/Python/Rust/SQL for semantic symbol lookup.
- **Code analysis** — import graphs, call hierarchies, blast radius, dead code, complexity, hotspots, cycles, importance, coupling, signal chains, layer violations, AST patterns, provenance, untested detection, PR risk, and unified coding context.
- **Doc indexing** — Markdown sections, links, glossary terms, code examples, broken link detection, tutorial path reconstruction.
- **Trust scoring** — memories linked to changed code lose confidence; stable linked code recovers trust.
- **HTTP server** — optional REST API for programmatic access to the Aurex domain model and code analysis.
- **Auto-detection** — the extension automatically captures decisions, bugfixes, and discoveries from assistant messages.
- **Claude Code bridge** — first-class CLI integration via MCP + hooks (`lapis claude-code install`).

## See Also

- [`CLAUDE_CODE.md`](CLAUDE_CODE.md) — Claude Code install, hook mapping, troubleshooting
- [`MCP.md`](MCP.md) — standalone MCP server (tools only)

- [`../skills/memory-layer/SKILL.md`](../skills/memory-layer/SKILL.md) — complete extension skill reference
- [`API.md`](API.md) — HTTP API and CLI reference
- [`TUTORIAL.md`](TUTORIAL.md) — step-by-step usage guide
- [`CONFIGURATION.md`](CONFIGURATION.md) — configuration options
- [`COMMANDS.md`](COMMANDS.md) — full CLI command reference

Files in this skill

  • API.md12.5 KB
  • ARCHITECTURE.md9.7 KB
  • ARCHITECTURE_MODULARIZATION.md23.9 KB
  • AUDIT-2026-06-16.md12.3 KB
  • CLAUDE_CODE.md10.7 KB
  • COMMANDS.md13.9 KB
  • CONFIGURATION.md8.9 KB
  • CONTRIBUTING.md658 B
  • DREAM_CYCLE.md2.3 KB
  • GITHUB_ISSUE_BREAKDOWN.md16.1 KB
  • HERMES.md4.3 KB
  • INDEX.md6.1 KB
  • JUDGMENT.md3.5 KB
  • MCP.md5 KB
  • MODULE_MAP.md8.8 KB
  • SKILL.md2.4 KB
  • TUTORIAL.md5.4 KB
  • code-indexing.md2.1 KB
  • decisions/crosshash-strategy.md13.7 KB
  • diagrams/lapis-architecture.html30.4 KB

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…