insights

Audits AI workflow health, vault structural integrity, and knowledge pipeline status.

Updated Nov 1, 2024
One-click install
npx skills add https://github.com/mlorentedev/dotfiles --skill insights-mlorentedev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: insights
Source: https://github.com/mlorentedev/dotfiles/tree/main/harness/skills/insights
Command: npx skills add https://github.com/mlorentedev/dotfiles --skill insights-mlorentedev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-assisted knowledge workflows silently degrade: agent memory files grow stale, vault links break, and bugfix or decision observations never get persisted into lessons or ADRs. This Skill runs a read-only audit that surfaces these issues before they compound. ## Core Features & Use Cases - MEMORY.md Health Check: Reports line count, last crystallization date, and flags entries older than 14 days or over 150 lines. - Vault Structural Audit: Runs vault health tests, detects unresolved wikilinks, orphan notes, and frontmatter compliance via Hive MCP and obs-cli. - Gap & Persistence Analysis (full mode): Inventories recent observations, finds bugfixes and decisions missing from docs/lessons, and verifies decisions were persisted to ADRs or patterns. - Use Case: Before starting a major sprint, run the full audit to confirm all prior decisions were persisted and the vault has no structural failures, then route issues to /crystallize or /vault-doctor. ## Quick Start Run the insights skill in quick mode to audit my MEMORY.md health, vault structure, and open backlog items.

Frequently Asked Questions about insights

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check the health of an AI agent memory file?

Run the insights audit, which reads MEMORY.md and reports line count, last crystallization date, and days since last update. It flags warnings when the file exceeds 150 lines or has not been crystallized in over 14 days.

How to detect broken wikilinks and orphan notes in Obsidian?

The audit runs vault_health via Hive MCP and, when the Obsidian GUI is running, executes obs-cli.sh unresolved and obs-cli.sh orphans. Results show unresolved link counts and orphan notes with OK or ACTION NEEDED status.

What is the difference between quick and full audit modes?

Quick mode covers MEMORY.md health, vault structural health, and a backlog snapshot in about two minutes. Full mode adds observation inventory, vault gap analysis, decision persistence checks, and pattern opportunity detection in about five minutes.

Does the insights audit modify any files in the vault?

No, the audit is strictly read-only and never modifies files. When it detects issues, it recommends running separate skills like /crystallize for knowledge gaps or /vault-doctor for structural problems.

When should I run a workflow health audit?

Run quick mode weekly as a maintenance habit or when session start shows vault warnings. Run full mode before starting a major sprint, after completing one, or when preparing for a crystallization pass.