# Pi Agent Embedding

> Pi agent embedding (`@mariozechner/pi-agent-core`)

- Skill: `romiluz13/pi-agent-embedding` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add romiluz13/pi-agent-embedding`
- Raw SKILL.md: https://api.skillmd.com/api/skills/romiluz13/pi-agent-embedding/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: romiluz13 (https://skillmd.com/u/romiluz13)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/romiluz13/pi-agent-embedding

---


# Pi agent embedding (`@mariozechner/pi-agent-core`)

## Grounding

1. `pi-mono/packages/agent/README.md` — Agent options, event tables, tool parallel vs sequential, `convertToLlm` / `transformContext` flow diagram.
2. `pi-mono/packages/coding-agent/examples/sdk/` — integration patterns (`README.md` index + numbered examples).
3. `pi-mono/packages/coding-agent/docs/sdk.md` — higher-level SDK documentation when present.

## Invariants

- Event ordering and barrier behavior (`message_end` before tool preflight) are specified in `pi-mono/packages/agent/README.md`; cite that file instead of restating from memory.

## Workflows

- **Pick integration depth**: README Quick Start vs full `examples/sdk/` depending on user needs; always point to concrete example filenames under `pi-mono/packages/coding-agent/examples/sdk/`.

## Anti-patterns

- Do not merge pi-coding-agent session semantics with core `Agent` without reading `agent-session.ts` when the question is about the CLI harness specifically — route CLI questions to `pi-cli-workspace` after citing boundaries from READMEs.