Date: 2026-04-13 Status: Ready to execute
- Create the repo layout for Codex skills, docs, scripts, and tests.
- Add maintainer docs:
- design spec
- implementation plan
- command coverage
- Add
.gitignore.
Create these skill folders:
skills/obsidian-vaultskills/obsidian-notesskills/obsidian-searchskills/obsidian-dailyskills/obsidian-tasksskills/obsidian-tagsskills/obsidian-bookmarksskills/obsidian-templatesskills/obsidian-pluginsskills/obsidian-ui
For each skill:
- Write
SKILL.mdwith:- trigger description in frontmatter
- Codex-specific output rules
- vault targeting guidance
- domain command reference
- domain-specific failure guidance
- Write
agents/openai.yamlwith:display_nameshort_descriptiondefault_prompt
- Document which current CLI commands map into v1 skills.
- Document intentionally deferred commands.
- Keep v1 claims aligned with actual coverage.
- Add
scripts/validate-skills.shto run the Codex validator on every skill. - Add
tests/smoke-test.shwith:- CLI presence check
- command transport preflight
- optional vault argument
- safe command execution without
eval
- Add
tests/README.mdmanual validation steps.
Review every skill for:
- frontmatter correctness
- accurate trigger descriptions
- command coverage within domain
- consistency of vault resolution rules
- no Claude-specific packaging or instructions
- no claim of raw-output passthrough that Codex cannot honor
- Run the skill validator.
- Run smoke tests in an environment where the Obsidian desktop app and CLI transport work.
- Record any environment-specific failures, especially WSL or redirector issues.