| name | mda-compat-codex-cli |
| description | MDA compatibility demonstration skill for Codex CLI. Compiled from compat/codex-cli/source.mda. Loading this skill confirms that an MDA-emitted SKILL.md is discovered and parsed by Codex CLI with the envelope subset of agentskills.io v1 fields honored. |
| allowed-tools | Read |
| metadata | {"mda":{"doc-id":"8e0a1c44-2b91-4d6e-9f7a-3c5d8e6f1a2b","title":"MDA Compat Demo (Codex CLI)","version":"1.0.0","tags":["compat","demo","codex-cli"]}} |
| integrity | {"algorithm":"sha256","digest":"sha256:a997097c76502fe6ebfa1d64a376b7737c3f3b37312589c76f1395fa2b61daa0"} |
MDA Compat Demo (Codex CLI)
This skill exists to verify Codex CLI SKILL.md compatibility from an MDA-compiled output.
Three things are true if Codex CLI discovers and loads this skill:
- The compiled
SKILL.md parses under Codex CLI's loader (built on agentskills.io v1).
- The frontmatter
name matches the package directory name (spec §06-2.1).
- The agentskills.io v1 envelope subset (
name, description, allowed-tools) survives the round-trip from .mda source to compiled output to runtime ingest.
Source: compat/codex-cli/source.mda. Compile: mda compile compat/codex-cli/source.mda --target SKILL.md --out-dir compat/codex-cli/build --integrity.
What this does NOT prove:
- That Codex CLI parses MDA-extended frontmatter under
metadata.mda.* (the OpenAI docs do not document any such behavior, and we do not require it).
- That signatures or integrity digests are verified at load time (no current runtime does that).
- That non-standard top-level fields observed in some shipped Codex skills (for example
triggers: arrays) are authored from MDA — they are not part of agentskills.io v1 and are out of MDA v1.0 scope.