Ontology KG

Typed knowledge graph for AI agent memory. Create entities, link relations, enforce constraints. JSONL append-only storage, zero dependencies. Use when agent...

Security Scan

The skill's description of an append-only JSONL ontology is reasonable, but the runtime instructions repeatedly call python3 scripts/ontology.py commands yet no scripts or install instructions are included — that mismatch is suspicious and needs clarification before use.

Purpose & Capability

The skill claims a zero-dependency JSONL store for a typed KG, which is coherent, but the SKILL.md examples and references assume a Python CLI (scripts/ontology.py). No such scripts or install steps are included in the package or manifest. That leaves a missing implementation: either the agent must implement the behavior itself or the package is incomplete.

Instruction Scope

Instructions tell the agent to create/read/relate/validate entities by running python scripts that read/write memory/ontology/graph.jsonl. The instructions reference local filesystem paths and document paths (which could point to arbitrary local files). While the described operations are appropriate for a memory/ontology tool, the SKILL.md grants implicit permission to read/write local files and to run arbitrary Python scripts — and since the scripts aren't provided, it's unclear what code would run.

Install Mechanism

There is no install spec (instruction-only), which is low risk in principle. However, the lack of included implementation while instructing use of specific scripts is an inconsistency: either an install or code files should be present or the SKILL.md should explain how the agent provides equivalent functionality.

Credentials

The skill requests no environment variables or credentials. The schema references secret_ref for credentials (encouraging not storing secrets directly), which is a reasonable safety constraint. There are no declared env/credential requirements that appear disproportionate.

Persistence & Privilege

always is false and model invocation is permitted (defaults). The skill does instruct writing to a local graph.jsonl (agent-local storage), which is expected for a memory tool. No indications it attempts to change other skills or gain elevated platform privileges.

Scan Findings in Context

[no_regex_findings] expected: The static regex scanner found nothing; this is expected because the skill is instruction-only and contains no executable code files to scan. Lack of findings does not prove safety — the SKILL.md references scripts that are not present.

What to consider before installing

Do not install or enable this skill until you confirm where the implementation lives. Questions to ask the publisher: (1) Where are scripts/ontology.py and any supporting code? The SKILL.md's CLI examples rely on that file. (2) If the code is provided separately, inspect it for filesystem and network access before running. (3) If you intend the agent to implement these semantics itself, understand the agent will create and modify memory/ontology/graph.jsonl (local files) — back up that folder and restrict file permissions. (4) Verify how credentials are handled: the schema requires secret_ref (good), but confirm the runtime never writes secrets into the JSONL. If the author cannot provide source code or a clear install path, treat the package as incomplete and avoid enabling autonomous invocation.

Ontology KG — 给你的 AI 一个结构化大脑

Story

你的 AI agent 记住了 200 条信息，全塞在一个 MEMORY.md 里。找东西靠 grep，关系靠脑补。

"Alice 负责哪个项目？" —— grep Alice，翻 10 条记录，猜。 "什么任务 block 了发布？" —— 没法回答，因为依赖关系没存。 "上次会议谁参加了？" —— 运气好能找到，运气不好淹在 200 行里。

问题不是记忆太少，是记忆没有结构。

Ontology KG 给 agent 的记忆加上类型、关系和约束。Person 有 name，Task 有 status 和 blocker，Project 有 owner。entity 之间用 relation 连接，约束自动校验。

存储是 JSONL append-only 文件，零依赖，git 友好。

Core Concept

Entity = { id, type, properties }
Relation = { from → rel_type → to }
Constraint = { type rules, relation rules, acyclic checks }

Types

Category	Types
People	Person, Organization
Work	Project, Task, Goal
Time	Event, Location
Info	Document, Message, Note
Resources	Account, Device, Credential
Meta	Action, Policy

Quick Start

# Init storage
mkdir -p memory/ontology
touch memory/ontology/graph.jsonl

# Create entities
python3 scripts/ontology.py create --type Person --props '{"name":"Alice"}'
python3 scripts/ontology.py create --type Project --props '{"name":"Website","status":"active"}'

# Link them
python3 scripts/ontology.py relate --from proj_001 --rel has_owner --to p_001

# Query
python3 scripts/ontology.py query --type Task --where '{"status":"open"}'
python3 scripts/ontology.py related --id proj_001 --rel has_task

# Validate all constraints
python3 scripts/ontology.py validate

Storage Format (JSONL)

{"op":"create","entity":{"id":"p_001","type":"Person","properties":{"name":"Alice"}}}
{"op":"relate","from":"proj_001","rel":"has_owner","to":"p_001"}

Append-only. Git-friendly. Migrate to SQLite for big graphs.

Key Constraints

• Credential cannot store secrets directly (must use secret_ref)
• blocks relation is acyclic (no circular dependencies)
• Event.end >= Event.start
• Task requires title + status

When to Use

Trigger	Action
"Remember that..."	Create/update entity
"What do I know about X?"	Query graph
"Link X to Y"	Create relation
"What depends on X?"	Traverse dependencies
"Show project status"	Aggregate by relations
Skills need shared state	Read/write entities

References

• references/schema.md — Full type definitions, relation types, constraints
• references/queries.md — Query patterns, traversal examples, aggregations

Credits

Inspired by oswalpalash/ontology. Optimized for practical agent use with streamlined schema and better defaults.