_shared

Provides shared SDD reference documents for persistence, conventions, and phase protocols.

Updated Aug 5, 2026
One-click install
npx skills add https://github.com/Alountk/Bloodbowl_Project --skill shared-alountk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: _shared
Source: https://github.com/Alountk/Bloodbowl_Project/tree/main/.opencode/skills/_shared
Command: npx skills add https://github.com/Alountk/Bloodbowl_Project --skill shared-alountk

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It centralizes common reference documentation for Spec-Driven Development (SDD) skills so that phase skills and orchestrators share one consistent set of conventions instead of duplicating instructions. ## Core Features & Use Cases - Persistence Contract: Defines artifact storage modes (engram, openspec, hybrid, none) and read/write behavior per mode. - Naming Conventions: Documents deterministic Engram topic keys and OpenSpec directory structures for SDD artifacts. - Phase Protocol: Supplies the common boilerplate every SDD phase sub-agent must follow, including skill loading, artifact retrieval, and return envelopes. - Use Case: When an orchestrator launches an SDD phase sub-agent, the sub-agent reads these shared references to know how to recover state, persist artifacts, and report results consistently. ## Quick Start Read the relevant shared reference file such as persistence-contract.md before implementing or invoking any SDD phase skill.

Frequently Asked Questions about _shared

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

FAQPage Schema
What is the _shared directory in SDD skills?▼

The _shared directory is a support package containing reference documents consumed by real SDD phase skills. It is explicitly not invokable as a skill itself and only provides common conventions and protocols.

How do SDD skills persist artifacts across sessions?▼

SDD skills persist artifacts using one of four modes: engram for memory-based storage, openspec for filesystem files, hybrid for both, or none for ephemeral output. The persistence contract defines read and write behavior for each mode.

What is the difference between engram and openspec storage modes?▼

Engram mode stores artifacts in local memory with topic-key upserts, supporting cross-session recovery but no history. OpenSpec mode writes files into the repository, giving git history and team shareability but no compaction survival.

Can I invoke the _shared skill directly?▼

No, the _shared package sets disable-model-invocation and user-invocable to false in its frontmatter. It exists only as reference documentation loaded by other SDD phase skills and orchestrators.

Why does mem_search require a follow-up retrieval call?▼

mem_search returns only truncated 300-character previews with observation IDs. The full artifact content requires a second call to mem_get_observation with the saved ID, which is mandatory for correct pipeline behavior.