메뉴
Extract canonical spec from existing agent or skill files
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Pre-action boundary checking — validates agent tool calls against declared capabilities and task contracts
Auto-detect project context and optimize harness — deactivate unused agents/skills, suggest missing experts, generate project profile
Multi-LLM adversarial consensus loop — 3+ LLMs compete to find flaws in designs/specs until unanimous agreement is reached
Monitor Claude Code releases and auto-generate GitHub issues for each new version
Execute OpenAI Codex CLI prompts and return results
YAML-based DAG workflow engine with topological execution and failure strategies
| name | omcustomcodex:takeover |
| description | Extract canonical spec from existing agent or skill files |
| scope | harness |
| user-invocable | true |
| argument-hint | <agent-name> |
Extract a canonical specification from an existing agent or skill file. Inspired by codespeak.dev's reverse compilation concept — deriving specs from existing implementations.
When an agent or skill has evolved organically without a formal spec, omcodex:takeover reverse-engineers a structured specification that captures its intent, invariants, workflow contract, and I/O contract.
/omcustomcodex:takeover <agent-name>
/omcustomcodex:takeover <skill-name>
1. Determine target type:
- Agent: read .codex/agents/<name>.md
- Skill: read .codex/skills/<name>/SKILL.md
2. Parse frontmatter (YAML metadata)
3. Parse body (markdown content)
Extract these components from the target:
| Component | Source | Description |
|---|---|---|
intent | Description field + first paragraph | One-line purpose statement |
invariants | Rules referenced, constraints mentioned | Things that must always be true |
workflow_contract | Workflow/stages sections | Input → processing → output steps |
io_contract | Tools field, input/output patterns | What the agent consumes and produces |
dependencies | Skills field, guide references | External knowledge required |
boundaries | Limitations field, disallowedTools | What the agent explicitly cannot do |
Output structured spec to .codex/specs/<name>.spec.md:
---
name: <name>
type: agent | skill
source: .codex/agents/<name>.md | .codex/skills/<name>/SKILL.md
generated: <ISO-8601 timestamp>
---
# Spec: <name>
## Intent
<one-line purpose>
## Invariants
- <rule or constraint that must always hold>
- ...
## Workflow Contract
### Input
<what the agent/skill receives>
### Processing
1. <step>
2. <step>
### Output
<what the agent/skill produces>
## I/O Contract
### Consumes
- Tools: [<tools used>]
- Files: [<files read>]
- MCP: [<MCP tools if any>]
### Produces
- Files: [<files created/modified>]
- Output: [<what is returned>]
## Dependencies
- Skills: [<referenced skills>]
- Guides: [<referenced guides>]
- Rules: [<rules enforced>]
## Boundaries
- <what the agent explicitly cannot do>
[Done] Spec extracted: .codex/specs/<name>.spec.md
├── Intent: <summary>
├── Invariants: <count> rules
├── Workflow: <step count> steps
└── Dependencies: <count> refs
.codex/)/omcustomcodex:takeover <name>/dev-refactor --spec for invariant-preserving refactoring