omos-init
Bootstrap baseline specs via $oh-my-openspec:omos-init — research the codebase, agree on capabilities, write openspec/specs/** describing what exists, critique and validate. Invoke via $oh-my-openspec:omos-init.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- omos-init
- description
- Bootstrap baseline specs via $oh-my-openspec:omos-init — research the codebase, agree on capabilities, write openspec/specs/** describing what exists, critique and validate. Invoke via $oh-my-openspec:omos-init.
03 SKILL.md 正文
Codex host adaptation (generated from skills/omos-init/SKILL.md — source of truth stays there):
- Dispatch roles with
spawn_agent/send_input/wait_agent/close_agent. Subagents are asynchronous (default up to 6 concurrent threads), so a wave is one batch ofspawn_agentcalls followed bywait_agentcollection — never poll.openspec_validateis provided by this plugin's bundled MCP server asmcp__openspec__openspec_validate— call it withcwdset to your session's project directory (thechangefilter optional); the VALID/INVALID/UNAVAILABLE/ERROR one-look report format is identical to OpenCode.- This skill is invoked as a
$oh-my-openspec:omos-initmention (or picked from /skills). There is no argument substitution: arguments arrive as plain text in the user's message after the mention — read them from there; when absent, follow this skill's elicitation steps.- Roles run from
~/.codex/agents/omos-*.toml. Ifspawn_agentreports an unknown agent, stop and tell the user to run$oh-my-openspec:omos-setuponce.
omos-init
Form the project's initial spec: baseline capability specs under
openspec/specs/ describing what the code does today. The orchestrator
dispatches, never drafts. You never run git write commands.
Flow
1. Preconditions
- Check the CLI: call
openspec_validateonce.UNAVAILABLE→ offer to install it now: runnpm install -g @fission-ai/openspec(suggestbrew install openspecon macOS when the user prefers it) — a global install is a system change, so proceed only on an explicit yes. Installed → re-check and continue. Declined → stop: unlike$oh-my-openspec:omos-plan, this command cannot degrade — bootstrapping specs without the mechanical gate is how baselines rot. - Check the project:
openspec/must exist. Missing → ask consent, runopenspec init --tools none(project artifacts only; the plugin already provides every OpenSpec skill). Declined → stop.
2. Baseline state
Read openspec/specs/.
- Empty → proceed to discovery.
- Non-empty → report existing capabilities; ask: extend (add only missing capabilities, never touch existing spec files) or stop. Extend mode: discovery skips covered capabilities.
3. Capability discovery
Dispatch omos-researcher subagents in parallel — one per suspected
area (entry points, packages, services; use the user's Focus areas: when
given). Each reports observable behaviors with path:line evidence.
Merge findings into a capability map:
<capability-name>: <one-line purpose> — <key evidence>. Propose names as
kebab-case nouns of what the system does, sized so each has 2–6
requirements.
4. Confirm the map
Present the map; interview until the user confirms add/remove/rename — at most 5 questions per round, each with a recommendation. A capability no researcher evidenced is dropped, not guessed.
5. Draft baselines
Dispatch omos-planner with MODE: baseline: the confirmed map plus the
research facts. It writes openspec/specs/<capability>/spec.md —
## Purpose (a real paragraph, not a stub), ## Requirements with
### Requirement: SHALL statements and #### Scenario: WHEN/THEN blocks
describing current, observed behavior only. Aspirations belong in
$oh-my-openspec:omos-plan, not here.
6. Critique loop (≤3 rounds)
Dispatch omos-critic in spec-review mode over the new files.
NEEDS-FIX→omos-plannerrevises the exact findings; repeat.CLEAN→ continue. StillNEEDS-FIXafter 3 rounds → present open findings, ask whether to accept or stop.- Mechanical INVALID lines from
openspec_validateare always fixed first.
7. Approval
Run openspec_validate (all specs and changes) and present: capability
list with requirement/scenario counts, gate output verbatim, every
UNVERIFIED line the critic flagged.
- Confirmed → final report: files, counts, gate result; next step —
future work goes through
$oh-my-openspec:omos-plan <idea>. Stop. - Rejected → collect the fix list, go to §5.
Guardrails
- Writes land only in
openspec/specs/— no changes/, no project code. - Existing spec files are never overwritten without per-file consent.
- Baseline = what the code does;
UNVERIFIEDbeats an invented SHALL. - Never commit, push, or stage anything.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.