opencode-agent-skills-md 1.0.1 → 1.1.0

package/.opencode/skills/test-skill/scripts/echo-args

@@ -1,14 +0,0 @@
-#!/bin/bash
-# Echo back arguments to test run_skill_script with args
-
-echo "Arguments received: $@"
-echo "Number of arguments: $#"
-
-if [ $# -eq 0 ]; then
-    echo "No arguments provided."
-else
-    echo "Arguments breakdown:"
-    for i in "$@"; do
-        echo "  - $i"
-    done
-fi

package/.opencode/skills/test-skill/scripts/greet

@@ -1,6 +0,0 @@
-#!/bin/bash
-# Simple greeting script to test run_skill_script
-
-echo "Hello from test-skill!"
-echo "Script executed successfully."
-echo "Current directory: $(pwd)"

package/AGENTS.md

@@ -1,43 +0,0 @@
-# Agent Guidelines
-
-## Commands
-- **Package manager:** pnpm.
-- **Install:** `pnpm install`
-- **Test:** `pnpm test` (runs `pnpm -r --no-bail test` for both packages, then the workspace contract test if all packages pass)
-- **Workspace contract test only:** `pnpm run test:workspace` (runs `node --import tsx --test tests/workspace.test.ts`)
-- **Single test file:** `pnpm -F opencode-agent-skills-md-core exec node --import tsx --test tests/<file>.test.ts` (or any file inside a package's `tests/`)
-- **Typecheck:** `pnpm run typecheck` (delegates to `pnpm -r run typecheck`)
-- **Build:** `pnpm run build` (delegates to `pnpm -r --workspace-concurrency=1 run build`)
-
-## Code Style
-- TypeScript is strict (`noUncheckedIndexedAccess` is on); check indexed access before using it.
-- ESM only; use `import`/`export` and `node:` builtins.
-- Runtime validation is manual (no Zod dependency).
-- Keep public functions documented with JSDoc when they are part of the plugin surface.
-- Prefer graceful fallbacks over hard failures for optional discovery and compaction hooks.
-- Validate user-supplied paths before reading files outside a skill root.
-
-## Repo Structure
-This is a pnpm workspace with two packages:
-- `packages/core/src/` — the portable, host-agnostic skills engine (`opencode-agent-skills-md-core`). Discovery, parsing, search, and the `SkillHostClient` / `SkillHostSession` boundary contracts.
-- `packages/opencode-agent-skills-md/src/` — the OpenCode plugin adapter (`opencode-agent-skills-md`). The four skill tools, the OpenCode host implementation, and the plugin factory.
-- Per-package tests live inside `packages/<pkg>/tests/`.
-- The repo root is a private workspace manifest (`package.json`) with no source of its own; legacy `src/`, root `tests/`, root `rolldown.config.js`, and root `tsconfig.build.json` were removed in favor of the per-package locations.
-
-## Skill Behavior
-- Discovery order is `.opencode/skills/`, `.claude/skills/`, `~/.config/opencode/skills/`, then `~/.claude/skills/`.
-- First match wins; duplicate skill names are ignored.
-- `use_skill` injects `SKILL.md` into context.
-- `read_skill_file` only reads inside the selected skill directory.
-- `run_skill_script` only runs executable files.
-- `OPENCODE_AGENT_SKILLS_SUPERPOWERS_MODE=true` enables the optional Superpowers bootstrap.
-
-## Task Tracking
-- Use **bd (beads)** for all issue/task tracking; do not create markdown TODO lists.
-- Commit `.beads/issues.jsonl` with code changes when issue state changes.
-- Run `bd sync` at the end of work sessions.
-- `.github/copilot-instructions.md` mirrors the bd workflow and is the repo-local source of truth.
-
-## Verification
-- Prefer `pnpm run typecheck` before `pnpm test` when you need a quick smoke check.
-- Keep docs aligned with the actual repo workflow; Bun references are stale and should not be reintroduced.

package/CHANGELOG.md

@@ -1,178 +0,0 @@
-# Changelog
-
-All notable changes to this project will be documented in this file.
-
-The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
-and this project attempts to adhere to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
-
-<!--
-## [${version}]
-### Added - for new features
-### Changed - for changes in existing functionality
-### Deprecated - for soon-to-be removed features
-### Removed - for now removed features
-### Fixed - for any bug fixes
-### Security - in case of vulnerabilities
-[${version}]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v${version}
--->
-
-## [Unreleased]
-
-### Changed
-
-- **Workspace split**: the repository now publishes two packages from a single pnpm workspace. The reusable, host-agnostic skills engine lives at [`opencode-agent-skills-md-core`](packages/core) and the OpenCode plugin adapter lives at [`opencode-agent-skills-md`](packages/opencode-agent-skills-md). The plugin package now consumes the core engine through a workspace dependency instead of relative source imports. The repo root is a private workspace manifest whose `build`, `test`, and `typecheck` scripts delegate to both packages via `pnpm -r`.
-
-### Added
-
-- The `opencode-agent-skills-md-core` package is published as a standalone ESM module so custom harnesses, CLIs, and test fixtures can embed the skills engine without pulling `@opencode-ai/plugin`. Consumers implement the `SkillHostClient` / `SkillHostSession` boundary contracts declared in `packages/core/src/types.ts` against their own host.
-
-### Removed
-
-- The legacy `opencode-agent-skills-md/core` subpath export is gone. Replace imports of `opencode-agent-skills-md/core` with `opencode-agent-skills-md-core`. The legacy root `src/`, root `tests/`, root `rolldown.config.js`, and root `tsconfig.build.json` have also been removed; each now lives inside the owning workspace package.
-
-## [0.7.0]
-
-### Added
-
-- Added `HF_ENDPOINT` support for configuring a Hugging Face mirror for embedding model downloads ([#38](https://github.com/MetalbolicX/opencode-agent-skills-md/pull/38), thanks [@pablon](https://github.com/pablon)).
-
-## [0.6.5]
-
-### Fixed
-
-- Fixed crash on macOS ARM64 (Apple Silicon) caused by `onnxruntime-node@1.21.0` segfault regression by pinning to 1.20.1 ([#31](https://github.com/MetalbolicX/opencode-agent-skills-md/issues/31), [microsoft/onnxruntime#24096](https://github.com/microsoft/onnxruntime/issues/24096))
-
-## [0.6.4]
-
-### Fixed
-
-- Fixed YAML frontmatter parsing for skills with multi-line descriptions (block scalar `|` and `>` syntax) by replacing custom parser with the `yaml` library
-- Added support for Claude's plugin v2 format in `installed_plugins.json`, which uses an array of installations per plugin instead of a single object
-
-### Changed
-
-- Claude Code plugin cache discovery now handles the new nested directory structure (`cache/<marketplace>/<plugin>/<version>/skills/`)
-
-## [0.6.3]
-
-### Changed
-
-- Improved skill evaluation prompt to hopefully prevent models from announcing "no skill required" style messages to users (who cannot see the hidden evaluation prompt)
-
-## [0.6.2]
-
-### Fixed
-
-- Skill validation now allows directory names to differ from the `name` in SKILL.md frontmatter. The frontmatter `name` is the canonical identifier, while directory names are for organization only. This aligns with the Anthropic Agent Skills spec.
-
-## [0.6.1]
-
-### Changed
-
-- Dynamic skill suggestions now track loaded skills per session and avoid re-suggesting already-loaded skills, reducing redundant prompts and context usage
-
-## [0.6.0]
-
-### Added
-
-- Semantic skill matching: after the initial skills list injection, subsequent messages are matched against skill descriptions using local embeddings
-- Added `@huggingface/transformers` dependency for local embedding generation (quantized all-MiniLM-L6-v2)
-- When a message matches available skills, injects a 3-step evaluation prompt (EVALUATE → DECIDE → ACTIVATE) to encourage skill loading (inspired by [@spences10](https://github.com/spences10)'s [blog post](https://scottspence.com/posts/how-to-make-claude-code-skills-activate-reliably))
-- Disk-cached embeddings for low-latency matching (~/.cache/opencode-agent-skills-md/)
-- Session cleanup on `session.deleted` event
-
-## [0.5.0]
-
-### Added
-
-- Added "Did you mean..." fuzzy matching suggestions when skill or script names are not found in all tools (`use_skill`, `read_skill_file`, `run_skill_script`, `get_available_skills`)
-
-### Changed
-
-- **BREAKING**: Renamed `find_skills` tool to `get_available_skills` for clearer intent
-- **Internal**: Reorganized codebase into separate modules (`claude.ts`, `skills.ts`, `tools.ts`, `utils.ts`, `superpowers.ts`) for better maintainability
-- **Internal**: Improved code quality by removing AI-generated comments and unnecessary code
-
-## [0.4.1]
-
-### Changed
-
-- Installation method now uses npm package via OpenCode config instead of git clone + symlink
-
-### Removed
-
-- Removed `INSTALL.md` (no longer needed with simplified installation)
-
-## [0.4.0]
-
-### Changed
-
-- Script discovery now recursively searches the entire skill directory (max depth 10) instead of only the root and `scripts/` subdirectory
-- Scripts are now identified by relative path (e.g., `tools/build.sh`) instead of base name
-- Renamed `skill_name` parameter to `skill` in `read_skill_file`, `run_skill_script`, and `use_skill` tools
-- Renamed `script_name` parameter to `script` in `run_skill_script` tool
-
-## [0.3.3]
-
-### Fixed
-
-- Fixed file and directory detection to properly handle symlinks by using `fs.stat`
-
-## [0.3.2]
-
-### Fixed
-
-- Preserve agent mode when injecting synthetic messages on session start
-
-## [0.3.1]
-
-### Fixed
-
-- Fixed unintended model switching when using skill tools by explicitly passing the current model during `noReply` operations (workaround for opencode issue #4475)
-
-## [0.3.0]
-
-### Added
-
-- Added file listing to `use_skill` output
-
-## [0.2.0]
-
-### Added
-
-- Added support for superpowers mode
-- Added release attestations
-
-## [0.1.0]
-
-### Added
-
-- Added `use_skill` tool to load skill content into context
-- Added `read_skill_file` tool to read supporting files from skill directories
-- Added `run_skill_script` tool to execute scripts from skill directories
-- Added `find_skills` tool to search and list available skills
-- Added multi-location skill discovery (project, user, and Claude-compatible locations)
-- Added Anthropic Agent Skills Spec v1.0 compliant frontmatter validation
-- Added automatic skills list injection on session start and after context compaction
-
-### New Contributors
-
-- Josh Thomas <josh@joshthomas.dev> (maintainer)
-
-[unreleased]: https://github.com/MetalbolicX/opencode-agent-skills-md/compare/v0.7.0...HEAD
-[0.1.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.1.0
-[0.2.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.2.0
-[0.3.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.3.0
-[0.3.1]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.3.1
-[0.3.2]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.3.2
-[0.3.3]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.3.3
-[0.4.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.4.0
-[0.4.1]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.4.1
-[0.5.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.5.0
-[0.6.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.0
-[0.6.1]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.1
-[0.6.2]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.2
-[0.6.3]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.3
-[0.6.4]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.4
-[0.6.5]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.6.5
-[0.7.0]: https://github.com/MetalbolicX/opencode-agent-skills-md/releases/tag/v0.7.0

package/Justfile

@@ -1,39 +0,0 @@
-set dotenv-load := true
-set unstable := true
-
-[private]
-default:
-    @just --list --list-submodules
-
-[private]
-fmt:
-    @just --fmt
-
-# Build both workspace packages (core engine + OpenCode plugin)
-build:
-    pnpm run build
-
-# Install dev plugin locally (project-scoped stub that re-exports the built bundle)
-install: build
-    mkdir -p .opencode/plugins
-    printf 'export { SkillsPlugin as default, SkillsPlugin } from "../../packages/opencode-agent-skills-md/dist/opencode/index.js";\n' > .opencode/plugins/skills.js
-
-# Uninstall local plugin copy
-uninstall:
-    rm -f .opencode/plugins/skills.js
-
-# Check if local plugin is installed
-status:
-    @ls -la .opencode/plugins/skills.js 2>/dev/null || echo "Not installed"
-
-# Run the full test suite (both packages + workspace contract test)
-test:
-    pnpm test
-
-# Run the workspace contract test in isolation
-test-workspace:
-    pnpm run test:workspace
-
-# Typecheck both workspace packages
-typecheck:
-    pnpm run typecheck

package/README.md

@@ -1,220 +0,0 @@
-# opencode-agent-skills-md
-
-<p align="center">
-  <a href="https://github.com/MetalbolicX/opencode-agent-skills-md/actions/workflows/release.yml"><img alt="release" src="https://img.shields.io/github/actions/workflow/status/MetalbolicX/opencode-agent-skills-md/release.yml?style=flat-square&logo=githubactions&label=release" /></a>
-  <a href="https://www.npmjs.com/package/opencode-agent-skills-md"><img alt="npm" src="https://img.shields.io/npm/v/opencode-agent-skills-md?style=flat-square&logo=npm" /></a>
-  <a href="LICENSE"><img alt="license" src="https://img.shields.io/github/license/MetalbolicX/opencode-agent-skills-md?style=flat-square" /></a>
-</p>
-
-<p align="center">Reusable Agent Skills engine plus an OpenCode plugin adapter, distributed as two workspace packages.</p>
-
-## Table of Contents
-
-- [Description](#description)
-- [Which Package?](#which-package)
-- [Features](#features)
-- [Installation](#installation)
-- [Usage](#usage)
-- [How it Works](#how-it-works)
-- [Examples](#examples)
-- [Contributing](#contributing)
-- [License](#license)
-
-## Description
-
-This repo publishes two packages:
-
-| Package | Purpose |
-|---------|---------|
-| [`opencode-agent-skills-md`](https://www.npmjs.com/package/opencode-agent-skills-md) | OpenCode plugin — the four skill tools (`use_skill`, `read_skill_file`, `run_skill_script`, `get_available_skills`) and the OpenCode host adapter. |
-| [`opencode-agent-skills-md-core`](packages/core) | Portable, host-agnostic skills engine: discovery, parsing, search, and the `SkillHostClient` / `SkillHostSession` boundary contracts. Zero dependency on `@opencode-ai/plugin`. |
-
-The core engine is the reusable engine; the OpenCode plugin is one concrete adapter built on top of it.
-
-## Which Package?
-
-Pick the package that matches your harness:
-
-- **You use OpenCode** → install [`opencode-agent-skills-md`](https://www.npmjs.com/package/opencode-agent-skills-md). It already implements the `SkillHostClient` boundary against the OpenCode SDK and ships the four tools ready to load.
-- **You build a custom harness, CLI, or test fixture** → install [`opencode-agent-skills-md-core`](packages/core). It is a standalone ESM package whose runtime dependencies exclude `@opencode-ai/plugin`. You provide your own `SkillHostClient` implementation and pass it to the tool factories of your choice.
-
-Both packages live in this repo as a pnpm workspace. From the repo root, `pnpm install` wires them together via the workspace link so the OpenCode plugin can resolve `opencode-agent-skills-md-core` by name during development.
-
-## Features
-
-- **Standardized Skill Discovery**: Finds skills from project and user locations, supporting both OpenCode and Claude skill directory layouts.
-- **Context Injection**: Loads `SKILL.md` content directly into the context window as synthetic, non-reply messages.
-- **Smart Keyword Matching**: Automatically monitors messages and uses lightweight keyword matching to invisibly prompt the agent to use relevant skills.
-- **Compaction Resilient**: Re-injects the list of loaded skills after session compaction events to ensure they remain available in long-running sessions.
-- **Script Execution**: Recursively finds and executes scripts (files with the executable bit set) within skill directories.
-- **Superpowers Mode**: Optional integration to automatically bootstrap the `using-superpowers` workflow.
-- **Reusable Engine**: The core package is host-agnostic — write a `SkillHostClient` for your harness and reuse the entire skills engine without pulling the OpenCode SDK.
-
-## Installation
-
-### Quick install (recommended)
-
-```bash
-npx opencode-agent-skills-md install
-```
-
-This registers the plugin in your global OpenCode config and verifies the installation. Restart OpenCode to activate.
-
-### CLI commands
-
-After installing, the following commands are available:
-
-- `oas install` — register the plugin in the global OpenCode config
-- `oas uninstall` — remove the plugin from the global OpenCode config
-- `oas status` — check whether the plugin is currently installed
-- `oas doctor` — validate the OpenCode configuration health
-
-### From npm (alternative)
-
-```bash
-npm install -g opencode-agent-skills-md
-```
-
-Then add the plugin to `~/.config/opencode/opencode.json`:
-
-```json
-{
-  "plugin": ["opencode-agent-skills-md"]
-}
-```
-
-To pin a specific version:
-
-```json
-{
-  "plugin": ["opencode-agent-skills-md@0.7.0"]
-}
-```
-
-### Manual configuration (fallback)
-
-If you prefer not to use the CLI, add the plugin entry manually to `~/.config/opencode/opencode.json`:
-
-```json
-{
-  "plugin": ["opencode-agent-skills-md"]
-}
-```
-
-Restart OpenCode after updating the config.
-
-### Custom harness (standalone engine)
-
-Install the core engine and implement `SkillHostClient` against your own host:
-
-```bash
-pnpm add opencode-agent-skills-md-core
-```
-
-```ts
-import {
-  discoverAllSkills,
-  parseSkillFile,
-  resolveSkill,
-  type Skill,
-  type SkillHostClient,
-  type SkillHostSession,
-} from "opencode-agent-skills-md-core";
-```
-
-The core package has zero runtime dependency on `@opencode-ai/plugin`, so it is the right entry point for custom harnesses, CLIs, and test fixtures. Implement the `SkillHostClient` interface declared in `packages/core/src/types.ts` and pass an instance to the tool factories of your choice.
-
-### Local development
-
-```bash
-git clone https://github.com/MetalbolicX/opencode-agent-skills-md
-cd opencode-agent-skills-md
-pnpm install
-pnpm run build    # builds both packages via `pnpm -r run build`
-```
-
-The OpenCode plugin bundle is emitted at `packages/opencode-agent-skills-md/dist/opencode/index.js`. Symlink it into your local OpenCode plugin directory:
-
-```bash
-mkdir -p ~/.config/opencode/plugin
-ln -sf "$(pwd)/packages/opencode-agent-skills-md/dist/opencode/index.js" ~/.config/opencode/plugin/skills.ts
-```
-
-## Usage
-
-### Skill Discovery Order
-
-The plugin discovers skills from these locations, in priority order (first match wins, preventing duplicates):
-
-1. `.opencode/skills/` (Project - OpenCode)
-2. `.claude/skills/` (Project - Claude)
-3. `~/.config/opencode/skills/` (User - OpenCode)
-4. `~/.claude/skills/` (User - Claude)
-
-### Tools
-
-Once loaded, use the tools directly from OpenCode:
-
-| Tool | Description |
-|------|-------------|
-| `use_skill` | Load a skill's `SKILL.md` into context. |
-| `read_skill_file` | Read a file from a skill directory (e.g., `references/rules.md`). |
-| `run_skill_script` | Execute a script from a skill directory. The script must be executable (`chmod +x`). |
-| `get_available_skills` | List available skills, optionally filtered by an ad-hoc query string. |
-
-### Superpowers Integration
-
-If you have the [Superpowers](https://github.com/obra/superpowers) skill installed, you can automatically inject its bootstrap prompt on session start:
-
-```bash
-export OPENCODE_AGENT_SKILLS_SUPERPOWERS_MODE=true
-opencode
-```
-
-## How it Works
-
-1. **Session Initialization**: On the first message, a complete list of discovered skills is injected into the context via `<available-skills>`.
-2. **Invisible Evaluation**: For subsequent messages, user text is tokenized and matched against skill names/descriptions. Matches trigger a hidden `<skill-evaluation-required>` block prompting the agent to invoke `use_skill` if appropriate.
-3. **Synthetic Injection**: Tool outputs are injected using `synthetic: true` and `noReply: true`, meaning they do not count as user messages and remain quietly in the context window.
-4. **Script Safety**: File reads (`read_skill_file`) strictly prevent path traversal outside the skill's root directory. Scripts (`run_skill_script`) skip common heavy directories like `node_modules` and `.git` and only execute files with the executable bit set.
-
-## Examples
-
-Load a skill into context:
-```text
-use_skill("brainstorming")
-```
-
-Read a supporting file from a skill:
-```text
-read_skill_file("brainstorming", "references/transformation-rules.md")
-```
-
-Run a script from a skill directory:
-```text
-run_skill_script("my-skill", "scripts/build.sh", ["--dry-run"])
-```
-
-List matching skills:
-```text
-get_available_skills({ query: "refactor" })
-```
-
-## Contributing
-
-Contributions are welcome. The codebase uses TypeScript and Rolldown for bundling. Two workspace packages live in this repo:
-
-- `packages/core/` — the portable skills engine.
-- `packages/opencode-agent-skills-md/` — the OpenCode plugin.
-
-Workspace commands at the repo root (`pnpm run build`, `pnpm test`, `pnpm run typecheck`) delegate to both packages via `pnpm -r`.
-
-Please run the following before opening a pull request:
-```bash
-pnpm run typecheck
-pnpm test
-```
-
-## License
-
-MIT. See the [LICENSE](LICENSE) file for details.

package/openspec/changes/archive/2026-06-14-skills-core-decouple/specs/core-decoupling/spec.md

@@ -1,74 +0,0 @@
-# Core Decoupling Specification
-
-## Purpose
-
-This spec defines the boundary between the portable core skills engine and the OpenCode-specific adapter in the `opencode-agent-skills-md` package. It guarantees that the core is reusable by any host while OpenCode users see no behavior change. This domain is introduced by the `skills-core-decouple` change; there is no prior baseline to delta against, so this file is the initial full definition.
-
-## Requirements
-
-### Requirement: Core Module Independence (R1)
-
-The `src/core/**` module graph SHALL contain zero runtime imports from `@opencode-ai/plugin`. A static import walk (AST or regex over the compiled dependency graph of `src/core/**`) SHALL fail the build if any such import is detected.
-
-### Requirement: Boundary Interface Location (R2)
-
-The interfaces `SkillHostClient` and `SkillHostSession` SHALL be declared in `src/core/types.ts`. The concrete OpenCode implementation of those interfaces SHALL exist only in `src/opencode/host.ts`. No other module SHALL contain a concrete implementation of either interface.
-
-### Requirement: Backward-Compatible Root Export (R3)
-
-Importing `opencode-agent-skills-md` (the package root) SHALL return the OpenCode plugin with the same shape and behavior as before this change. A smoke test that loads the package and resolves the four tool names (`use_skill`, `read_skill_file`, `run_skill_script`, `get_available_skills`) SHALL pass.
-
-### Requirement: Framework-Agnostic Subpath Export (R4)
-
-Importing `opencode-agent-skills-md/core` SHALL resolve to the core modules only. A smoke import of the subpath SHALL NOT trigger any side effect, top-level await, or transitive import from `@opencode-ai/plugin`.
-
-### Requirement: Discovery Semantics Preservation (R5)
-
-When the OpenCode adapter delegates to the core's `discoverAllSkills()`, the adapter SHALL observe the same four-location priority and first-match-wins semantics defined by the existing skill behavior. An integration test that mirrors `tests/integration/*` SHALL pass against the new boundary.
-
-### Requirement: Public Surface Freeze (R6)
-
-The four tool names, their parameter shapes, and their user-visible error messages SHALL remain unchanged. Existing tests under `tests/integration/*` and `tests/e2e/*` SHALL pass without modification.
-
-## Scenarios
-
-### Scenario: core is decoupled from the OpenCode SDK
-
-- GIVEN the `src/core/**` module graph is fully composed
-- WHEN a static import walk scans every file in that graph for `from "@opencode-ai/plugin"`
-- THEN zero matches are reported
-- AND the test `tests/core/agnostic.test.ts` passes
-
-### Scenario: opencode host is the only concrete implementation
-
-- GIVEN the interfaces `SkillHostClient` and `SkillHostSession` exist in `src/core/types.ts`
-- WHEN the codebase is searched for classes or factories that implement either interface
-- THEN exactly one match exists, located in `src/opencode/host.ts`
-
-### Scenario: root export still loads the OpenCode plugin
-
-- GIVEN a consumer imports the package via its root export
-- WHEN the consumer resolves the four tool names from the returned plugin object
-- THEN all four tool names (`use_skill`, `read_skill_file`, `run_skill_script`, `get_available_skills`) are present
-- AND the test-skill load assertion in `tests/integration/startup-smoke.test.ts` continues to pass
-
-### Scenario: subpath export does not pull in the OpenCode SDK
-
-- GIVEN a consumer imports `opencode-agent-skills-md/core` in a fresh process
-- WHEN the import resolves and the module graph is recorded
-- THEN no module under `node_modules/@opencode-ai/plugin` appears in the resolved graph
-- AND no side effect from the SDK runs at import time
-
-### Scenario: discovery priority and first-match-wins are preserved
-
-- GIVEN skill fixtures exist under `.opencode/skills/`, `.claude/skills/`, `~/.config/opencode/skills/`, and `~/.claude/skills/`
-- WHEN the OpenCode adapter calls `discoverAllSkills()`
-- THEN the returned list follows the four-location priority
-- AND duplicate skill names resolve to the first matching location
-- AND the new integration test in `tests/core/` passes
-
-### Scenario: public tool surface is unchanged
-
-- GIVEN the refactor is complete and `pnpm run typecheck` is clean
-- WHEN `pnpm test` runs the existing `tests/integration/*` and `tests/e2e/*` suites
-- THEN all existing assertions on tool names, parameter shapes, and error messages continue to pass without modification

package/openspec/changes/archive/2026-06-14-skills-core-decouple/tasks.md

@@ -1,64 +0,0 @@
-# Tasks: Skills Core Decouple
-
-## Review Workload Forecast
-
-| Field | Value |
-|-------|-------|
-| Estimated changed lines | PR1≈350; PR2≈250; PR3≈150; aggregate≈750 |
-| 400-line budget risk | Low per PR |
-| Chained PRs recommended | Yes |
-| Suggested split | PR1 core-extraction → PR2 opencode-adapter → PR3 subpath-exports |
-| Delivery strategy | force-chained |
-| Chain strategy | stacked-to-main |
-
-Decision needed before apply: No
-Chained PRs recommended: Yes
-Chain strategy: stacked-to-main
-400-line budget risk: Low
-
-### Suggested Work Units
-
-| Unit | Goal | Likely PR | Notes |
-|------|------|-----------|-------|
-| 1 | Extract portable core with import guard | PR 1 | Base=main; merge after `pnpm run typecheck` + `pnpm test` |
-| 2 | Add OpenCode host and thin adapter | PR 2 | Base=main after PR1 merge; keep root behavior unchanged |
-| 3 | Publish subpath exports and built smoke tests | PR 3 | Base=main after PR2 merge; verify build + e2e |
-
-## Phase 1: PR1 — Core Extraction
-
-- [x] 1.1 Capture a green baseline with `pnpm run typecheck` and `pnpm test` before moving files.
-- [x] 1.2 RED: add `tests/core/agnostic.test.ts` with a static walk over `dist/core/**` for zero `@opencode-ai/plugin` imports.
-- [x] 1.3 Move host-agnostic code from `src/utils.ts` and `src/skills.ts` into `src/core/{types,parse,discovery,scripts,match,content}.ts`.
-- [x] 1.4 Add `src/core/index.ts`; update `src/utils.ts` and `src/skills.ts` to re-export or consume core without behavior changes.
-- [x] 1.5 Update `rolldown.config.js` for a temporary `dist/core` entry while preserving the current plugin build. *(Superseded by PR3 3.2 — multi-entry build emits both `dist/opencode/index.js` and `dist/core/index.js` plus the legacy `dist/plugin.mjs` alias.)*
-- [x] 1.6 GREEN gate: `pnpm run typecheck` and `pnpm test` pass with the new core boundary.
-
-## Phase 2: PR2 — OpenCode Adapter
-
-- [x] 2.1 Add `src/opencode/host.ts` as the only concrete `SkillHostClient` and `SkillHostSession` implementation.
-- [x] 2.2 Add `src/opencode/tools.ts` and rewrite tool factories to compose core logic through the host.
-- [x] 2.3 Move plugin composition to `src/opencode/plugin.ts`; keep `src/plugin.ts` as a thin compatibility shim until PR3.
-- [x] 2.4 Move `src/superpowers.ts` to `src/opencode/superpowers.ts` and wire `src/opencode/index.ts` as the adapter entry.
-- [x] 2.5 Verify R5/R6 by running `tests/integration/plugin.test.ts` and `tests/e2e/startup-smoke.test.ts` unchanged.
-- [x] 2.6 GREEN gate: `pnpm run typecheck` and `pnpm test` pass before merge.
-
-## Phase 3: PR3 — Subpath Exports (public surface change)
-
-- [x] 3.1 Add `exports` map to `package.json` with `.` and `./core` subpaths (each with `types` + `import` + `default` conditions) and point `main` at `dist/opencode/index.js`.
-- [x] 3.2 Rewrite `rolldown.config.js` to multi-entry emit `dist/opencode/index.js` and `dist/core/index.js`; add `tsconfig.build.json` for `tsc --emitDeclarationOnly` so `.d.ts` files travel with the JS chunks.
-- [x] 3.3 Add `tests/core/subpath.test.ts` — resolves `opencode-agent-skills-md/core` via the package's exports field, asserts the portable API is exported, and walks the built chunk to prove no `@opencode-ai/plugin` reference leaks in.
-- [x] 3.4 Add `tests/opencode/subpath.test.ts` — resolves `opencode-agent-skills-md` (root) via the exports field, asserts the default export + `SkillsPlugin` are the same factory function, and proves module load is side-effect safe.
-- [x] 3.5 Add `pretest` hook that runs `pnpm run build` first so the subpath tests always run against fresh built artifacts.
-- [x] 3.6 Remove `src/plugin.ts` compat shim (now redundant); build entry is `src/opencode/index.ts` directly. A legacy `dist/plugin.mjs` alias is preserved via a third rolldown entry so any downstream consumer still importing the old path keeps working.
-- [x] 3.7 Update `README.md` with a "Programmatic subpath exports" section documenting both subpaths, the intended consumer (harness authors), and a code example using the core.
-- [x] 3.8 GREEN gate (final): `pnpm run typecheck` + `pnpm test` + `pnpm run build` all green; the new subpath tests pass against the freshly built artifacts.
-- [x] 3.9 Update `openspec/changes/skills-core-decouple/tasks.md` to mark all PR3 tasks done and persist the final apply-progress to engram.
-
-### PR3 Outcome
-
-- `pnpm run typecheck` — clean (0 errors)
-- `pnpm test` — 42/42 (21 unit + 3 agnostic + 3 core-subpath + 8 opencode-host + 3 opencode-subpath + 3 integration + 1 e2e)
-- `pnpm run build` — emits `dist/opencode/{index.js,index.d.ts,...}`, `dist/core/{index.js,index.d.ts,...}`, and the legacy `dist/plugin.mjs` alias
-- The `src/opencode/tools.ts` factory consts needed explicit `: ReturnType<typeof tool>` annotations to break a Zod-shape leak in the inferred return type that broke `tsc --emitDeclarationOnly` with TS2883 (non-portable inferred type). Internal change only — no behavior delta, no public API change.
-- The `package.json#exports` block needed a `default` condition on each subpath in addition to `import` so Node 24 can resolve the package via self-reference when called from a CJS context (`require.resolve`).
-- The integration test (`tests/integration/plugin.test.ts`) and the e2e test (`tests/e2e/startup-smoke.test.ts`) had their import targets updated to the new entry points (`src/opencode` and `dist/opencode/index.js` respectively).