$npx -y skills add jmstar85/oh-my-githubcopilot --skill planStrategic planning with optional interview workflow. Activate when user says: plan this, plan the, let's plan, make a plan, how should we approach.
| 1 | # Plan |
| 2 | |
| 3 | Creates comprehensive, actionable work plans through intelligent interaction. Auto-detects whether to interview (broad requests) or plan directly (detailed requests). |
| 4 | |
| 5 | ## Modes |
| 6 | |
| 7 | | Mode | Trigger | Behavior | |
| 8 | |------|---------|----------| |
| 9 | | Interview | Default for broad requests | Interactive requirements gathering | |
| 10 | | Direct | `--direct`, or detailed request | Skip interview, generate plan directly | |
| 11 | | Consensus | `--consensus`, "ralplan" | Planner → Architect → Critic loop | |
| 12 | | Review | `--review` | Critic evaluation of existing plan | |
| 13 | |
| 14 | ## Interactive Hook Protocol |
| 15 | |
| 16 | **MANDATORY**: Use `vscode_askQuestions` for ALL user-facing decision points in this skill (when available). |
| 17 | If `vscode_askQuestions` is NOT available (e.g., Copilot CLI), present numbered options in markdown and ask the user to respond with a number or freeform text. |
| 18 | |
| 19 | ### When to Fire Hooks |
| 20 | | Trigger Point | Question Type | |
| 21 | |---------------|---------------| |
| 22 | | Interview mode: each question round | Scope/preference/constraint question with options | |
| 23 | | Interview mode: readiness check | "Ready to generate plan?" gate | |
| 24 | | Consensus mode: plan trade-offs | Present design options with pros/cons | |
| 25 | | Consensus mode: critic rejection | Show rejection reasons, ask for direction | |
| 26 | | All modes: plan approval | Final plan review before execution | |
| 27 | |
| 28 | ## Interview Mode (broad/vague requests) |
| 29 | 1. Classify request: broad triggers interview |
| 30 | 2. **HOOK: Ask ONE focused question via `vscode_askQuestions`** for preferences, scope, constraints |
| 31 | - Provide 3-5 contextual options derived from codebase analysis |
| 32 | - Always include freeform input (`allowFreeformInput: true`) |
| 33 | 3. Gather codebase facts via @explore BEFORE asking user |
| 34 | 4. Consult @analyst for hidden requirements |
| 35 | 5. **HOOK: Readiness gate** — ask user if ready to generate plan: |
| 36 | ``` |
| 37 | header: "plan-readiness" |
| 38 | question: "I've gathered enough context. Ready to generate the plan?" |
| 39 | options: [ |
| 40 | { label: "Yes, generate the plan", recommended: true }, |
| 41 | { label: "I have more requirements to add" }, |
| 42 | { label: "Show me what you've gathered so far" } |
| 43 | ] |
| 44 | ``` |
| 45 | 6. Create plan when user signals readiness |
| 46 | |
| 47 | ## Direct Mode (detailed requests) |
| 48 | 1. Optional brief @analyst consultation |
| 49 | 2. Generate comprehensive work plan immediately |
| 50 | |
| 51 | ## Consensus Mode (`--consensus` / "ralplan") |
| 52 | 1. @planner creates initial plan with RALPLAN-DR summary (Principles, Decision Drivers, Options) |
| 53 | 2. **HOOK: Present design trade-offs** when viable options have significant trade-offs: |
| 54 | ``` |
| 55 | header: "design-trade-off" |
| 56 | question: "The planner identified competing approaches. Which direction?" |
| 57 | options: [derived from RALPLAN-DR viable options with pros/cons in descriptions] |
| 58 | ``` |
| 59 | 3. @architect reviews for architectural soundness (sequential, NOT parallel with critic) |
| 60 | 4. @critic evaluates quality criteria (after architect completes) |
| 61 | 5. If critic rejects: **HOOK** — show rejection reasons, ask user for direction: |
| 62 | ``` |
| 63 | header: "critic-rejection" |
| 64 | question: "Critic flagged issues: [summary]. How to proceed?" |
| 65 | options: [ |
| 66 | { label: "Address all issues", recommended: true }, |
| 67 | { label: "Address critical issues only" }, |
| 68 | { label: "Override — accept plan as-is" }, |
| 69 | { label: "Start over with different approach" } |
| 70 | ] |
| 71 | ``` |
| 72 | 6. Re-review loop (max 5 iterations) |
| 73 | 7. Final plan includes ADR (Decision, Drivers, Alternatives, Why chosen, Consequences) |
| 74 | |
| 75 | ## Review Mode (`--review`) |
| 76 | 1. Read plan from `.omg/plans/` |
| 77 | 2. @critic evaluates |
| 78 | 3. Return verdict: APPROVED / REVISE / REJECT |
| 79 | |
| 80 | ## Plan Approval Hook (all modes) |
| 81 | **HOOK**: After plan generation, present plan summary for user approval: |
| 82 | ``` |
| 83 | header: "plan-approval" |
| 84 | question: "[N] tasks identified. Review the plan and choose next step:" |
| 85 | options: [ |
| 86 | { label: "Approve & execute with autopilot", recommended: true }, |
| 87 | { label: "Approve & execute with ralph" }, |
| 88 | { label: "Approve & execute with team" }, |
| 89 | { label: "Revise — I have feedback" }, |
| 90 | { label: "Save plan only — don't execute yet" } |
| 91 | ] |
| 92 | ``` |
| 93 | |
| 94 | ## Output |
| 95 | Plans saved to `.omg/plans/`. Include: |
| 96 | - Requirements Summary |
| 97 | - Testable Acceptance Criteria |
| 98 | - Implementation Steps (with file references) |
| 99 | - Risks and Mitigations |
| 100 | - Verification Steps |