원클릭으로
darwin-plan-template
Structured plan template for the Architect agent. Use when creating infrastructure or code change plans.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Structured plan template for the Architect agent. Use when creating infrastructure or code change plans.
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
Safety rules for modifying Dockerfiles and Containerfiles. Use when editing container build files.
MR/PR context gathering and diagnostic reporting guidelines. Activates when working on events that reference MRs/PRs or when reporting investigation findings.
Team coordination for implement and test modes. Developer and QE use team_send_results for final reports, team_huddle for mid-task questions to FRIDAY.
GitOps workflow rules for modifying infrastructure via git. Use when cloning repos, modifying Helm values, committing, pushing, or verifying deployments.
MR/PR lifecycle operations -- pipeline check, retest, merge, conflict reporting. Extends darwin-gitlab-ops.
Report findings and status updates to the Darwin FRIDAY. Use team_send_results for final reports (task modes). In message mode, use team_send_message instead.
| name | darwin-plan-template |
| description | Structured plan template for the Architect agent. Use when creating infrastructure or code change plans. |
| roles | ["architect"] |
| modes | ["plan"] |
Use this structure for all plans:
---
plan: [Action] [Target]
service: [name]
repository: [repo URL]
path: [helm path or source path]
domain: [CLEAR|COMPLICATED|COMPLEX]
risk: [low|medium|high]
steps:
- id: 1
agent: [agent]
mode: [mode]
summary: [short description]
status: pending
- id: 2
agent: [agent]
mode: [mode]
summary: [short description]
status: pending
---
# Plan: [Action] [Target]
## Reason
[Why this change is needed, based on evidence]
## Steps
1. Specific action with implementation details
2. Specific action with implementation details
## UX Considerations (include when plan touches frontend/UI)
- User goal: [what the user is trying to accomplish]
- Current friction: [what's wrong with the current experience]
- Interaction pattern: [expand/collapse, modal, inline edit, etc.]
- States: [loading, empty, error -- how each is handled]
## Risk Assessment
- Risk level: [low/medium/high]
- Rollback: [how to undo]
## Verification
- [How will we know the change worked?]
- [What metric or signal confirms success?]
The frontmatter YAML header is machine-readable by FRIDAY and executing agents (Developer, QE).
The status field starts as pending and the executing team updates it to in_progress, completed, or failed as they work through the steps.
The Markdown body below the frontmatter contains the full human-readable details for each step.
Assign each step in the frontmatter steps: array to an agent and mode:
sysAdmin -- Kubernetes and GitOps operations
investigate -- Read-only cluster and service inspection, remote cluster accessexecute -- GitOps mutations (Helm value changes, commit, push)rollback -- Revert last GitOps change, verify CD syncdeveloper -- Code implementation agent
investigate -- Read-only: check MR/PR status, code inspectionexecute -- Single write actions: post comment, merge MR, tag releaseimplement -- Code changes. FRIDAY dispatches Developer first, then QE for verification.qe -- Quality verification agent
test -- Run tests, verify deployments via browser (Playwright MCP or repo Playwright tests)investigate -- Read-only test status checksarchitect -- Planning and review (use sparingly, avoid self-referential loops)
review -- Code/MR/PR review with severity findingsanalyze -- Information gathering and status reportHow to write steps for developer and QE:
developer with mode: implement,
and verification steps to qe with mode: test. FRIDAY dispatches them sequentially.developer with mode: investigate.developer with mode: execute.For COMPLICATED plans with multiple options, present options WITHOUT step assignments in the frontmatter.
Only the selected option's execution steps get agent and mode fields.