@dzhechkov/adapter-agents-md

The AGENTS.md platform adapter for DZ Harness Hub — flattens canonical skills into a single, root-level AGENTS.md.

dz init --target agents-md --select design-thinking

What it emits

AGENTS.md is the emerging cross-tool convention read by ~15 agents (Cursor, Zed, Warp, Aider, goose, Gemini CLI, RooCode, Kilo, Junie, Trae, Augment, Devin, pi, Windsurf, …). It is one root-level, plain-Markdown file — no YAML frontmatter, no per-skill directory:

AGENTS.md   # a plain-Markdown "## <skill>" section per selected skill, inside a dz-managed fenced block

Unlike the skill-tree adapters (one directory per skill) and unlike copilot (one instruction file per skill), one AGENTS.md holds all selected skills.

Intentionally lossy — and merge-not-overwrite

Like copilot, this is a lossy adapter. AGENTS.md loses:

• YAML frontmatter — the section is plain Markdown;
• progressive disclosure — the whole section is always-on;
• per-skill file boundaries — every skill collapses into one file;
• assets — scripts//references/ are not carried.

Because AGENTS.md is often hand-authored, aggregation into the real file (at the operations layer) uses @dzhechkov/core's mergeAgentsMd, which owns only a fenced block:

<!-- dz:skills BEGIN (managed by dz — do not edit) -->
…skill sections…
<!-- dz:skills END -->

Everything outside the fence is preserved byte-for-byte. The merge is idempotent.

Per the @dzhechkov/core Adapter contract the loss surfaces as a warning; under strict the adapter throws. The canonical skill stays the lossless source of truth — recompiling to claude is still full fidelity.