Skills interop

Agent Knowledge and Agent Skills should have clear responsibilities.

• Agent Knowledge stores facts, sources, compiled artifacts, status, review, and traceability records.
• Agent Skills store procedures, scripts, tool calls, prompt templates, and maintenance methods.

A healthy ecosystem should use Skills to maintain Knowledge, not hide concrete customer, brand, research, or organizational knowledge inside global Skills.

Layer model

A maintenance Skill can create, compile, lint, evaluate, and publish knowledge packs. A compatible client still treats knowledge as data during discovery and activation, and must not execute content found inside a knowledge pack.

Companion Skill

We recommend capturing maintenance workflows in a companion Skill such as agent-knowledge-maintainer.

It can provide:

• creating a knowledge pack
• importing sources and normalizing metadata
• compiling sources/ -> wiki/ -> compiled/ + indexes/
• running health and citation checks
• running discovery, context, and answer evals
• generating version snapshots and changelogs

These capabilities belong to the procedural layer and should live in a Skill, client command, CI, or external tool. The knowledge pack can store schemas, eval cases, run records, and sample data, but should not require clients to execute bundled scripts.

Script boundary

If a companion Skill uses scripts, scripts should follow the maintenance script contract:

• write operations support --dry-run
• output machine-readable JSON
• diagnostics go to stderr
• dependencies and runners are pinned
• network access and credential use are declared explicitly

What stays out of the pack core

The following may exist in Skills or toolchains, but should not become required Agent Knowledge protocol:

• a scripts/ directory
• a specific LLM, editor, vector store, or graph database
• a specific package manager
• concrete importers, crawlers, or converters
• proprietary commands for one client

The portable Agent Knowledge unit remains a plain directory with Markdown and JSON artifacts.

Interop principles

• A Skill can write a knowledge pack, but a knowledge pack cannot require a client to execute a Skill.
• Skill output must leave runs/ records explaining what was read, what changed, and why review is needed.
• The pack's status, trust, and grounding remain determined by pack metadata and review results.
• A client may call a maintenance Skill, but runtime answers should read maintained knowledge artifacts through a resolver.