Menu
Use `design-conventions` to get Project Conventions when deciding on names, job vocabulary, or skill terms to avoid repo-specific ambiguity.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Use `author-caveman` to write Caveman English.
Use `author-agent` to get OpenCode Agent Authoring when writing or reviewing OpenCode agent markdown.
Use `author-command` to get OpenCode Command Authoring when writing or reviewing OpenCode command markdown.
Use `author-rules` to write `AGENTS.md` files.
Use `author-skill` to get Skill File Authoring when you need to review/write skill files for agents.
Use `design-prd` to get Product Requirements when planning any feature or to understand project business requirements, user roles, and success criteria.
| name | design-conventions |
| description | Use `design-conventions` to get Project Conventions when deciding on names, job vocabulary, or skill terms to avoid repo-specific ambiguity. |
session.yml.job_name..agents/jobs/concepts/ before planning.plan.md under .agents/jobs/drafts/{name}/.job_name and lifecycle directory.concepts, drafts, assist, executing, facilitate, review, shelved.shelved.src/skills/*/SKILL.md written to the user skill store.assist, auto, research, design.temp_.Purpose: Keep job state and filesystem paths aligned.
Pattern: Use exact lifecycle names: concepts, drafts, assist, executing, facilitate, review, shelved. Do not invent new status words.
Purpose: Make jobs discoverable and movable across states.
Pattern: Use .agents/jobs/{status}/{job_name}/ with plan.md, criteria.yml, solution.md, and session.yml as needed.
Purpose: Preserve one job identity across session and title updates.
Pattern: Store session_id in session.yml; sync session title to job_name when lifecycle changes.
Purpose: Separate user commands from internal tool names.
Pattern: User commands use kebab-case job-* and document-*; tool names use autocode_*; internal agents use temp_*, execute_*, query_*, document_*, auto_*.
Purpose: Keep bundled skills deterministic.
Pattern: Source skills live in src/skills/{name}/SKILL.md; generated output is written to ~/.agents/skills/autocode/{name}/SKILL.md (or XDG config equivalent).