hyperplan
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', 'adversarial plan', 'hostile planning', 'cross-critique plan', '하이퍼플랜', '적대적 계획', '교차 비평'.
Profile →npx skills add code-yeongyu/oh-my-openagent --skill hyperplan --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
The skill acts as a Lead orchestrator for a 5-member adversarial team. It enforces a ruthless cross-critique debate among the specified categories (unspecified-low, unspecified-high, deep, ultrabrain, artistry). It distills only defensible insights from the attacks and mandatorily hands the distilled insight bundle to the plan agent for executable plan formalization.
How it works
- On load, it must say "HYPERPLAN MODE ENABLED!" as the first action.
- It requires preconditions: team_mode tools available, running as a lead-eligible agent (not a planner), and being in the main session.
- It defines five adversarial members with system prompts for skepticism, validation, research, architecture, and creativity, each performing independent analysis that yields 3-7 findings (each ≤3 sentences) without synthesis.
- Phase 1: Acknowledge, capture request, and create a Phase 7 todo including a mandatory Phase 6 plan handoff.
- Phase 2: Spawn team with team_create using a detailed inline_spec. Capture teamRunId for subsequent calls; retry without the researcher if deep is unavailable.
- Phase 3: Round 1 independently sends identical prompts to all five members to produce findings only.
- Phase 5: Phase 5 is the distillation phase where surviving insights are distilled; Phase 6 requires the plan agent to formalize an executable plan.
- Handoff to plan agent in Phase 6 is mandatory.
When to use it
Use when planning requires maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggered by keywords: hyperplan, hpp, /hyperplan, adversarial plan, hostile planning, cross-critique plan, 하이퍼플랜, 적대적 계획, 교차 비평.
What it can touch
Tools declared: claude-code, codex, cursor are available for execution within the skill workflow. The skill interacts with these tools through the team orchestration framework.
Caveats
Preconditions must be satisfied: team_mode.enabled must be true in the config, and you must be the lead-eligible agent. If deep is unavailable, the deep member may be omitted per instruction, but the other four must be present. The license is NOASSERTION.
The workflow explicitly prohibits providing synthesized results before the Phase 5 distillation and Phase 6 handoff to the plan agent. The skill emphasizes defending defensible insights and rejecting unfounded claims without explicit evidence.
# HYPERPLAN — Adversarial Multi-Agent Planning > **MANDATORY**: First action when this skill loads — say "HYPERPLAN MODE ENABLED!" so the user knows orchestration started. ## WHAT THIS IS You (the orchestrator) become the **Lead** of a 5-member adversarial team. The 5 members are **maximally hostile** to each other — they attack each other's findings ruthlessly. You then synthesize only the **defensible insights** that survived the attacks into a work plan. This is not consensus building. This is intellectual combat. Weakness gets exposed. Lazy thinking gets eviscerated. Only what survives the gauntlet makes it into the plan. ## HARD PRECONDITIONS Before starting, verify: 1. **`team_*` tools must be available.** If they are not, STOP and tell the user: > "Hyperplan requires team-mode. Set `team_mode.enabled: true` in `~/.config/opencode/oh-my-opencode.jsonc` and restart opencode, then retry." 2. **You are running as `sisyphus` (or another lead-eligible agent).** If you are running as a planner (`prometheus`, `plan`), this skill is the wrong tool — direct the user to use `/start-work` instead. 3. **You are in the main session** (not a background subagent). Hyperplan only works as a
- WHAT THIS IS
- HARD PRECONDITIONS
- THE 5 ADVERSARIAL MEMBERS — RnR & CHARACTERISTICS
- CATEGORY CHARACTERISTICS REFERENCE
- MEMBER 1: skeptic (category: unspecified-low)
- MEMBER 2: validator (category: unspecified-high)
- MEMBER 3: researcher (category: deep)
- MEMBER 4: architect (category: ultrabrain)
- MEMBER 5: creative (category: artistry)
- EXECUTION WORKFLOW
- Phase 0: Acknowledge and capture the request
- Phase 1: Spawn the adversarial team
- Phase 2: Round 1 — Independent analysis
- Phase 3: Round 2 — Cross-attack
What does the hyperplan skill do?
Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', 'adversarial plan', 'hostile planning', 'cross-critique plan', '하이퍼플랜', '적대적 계획', '교차 비평'.
How do I install it?
Run `npx skills add code-yeongyu/oh-my-openagent --skill hyperplan --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From code-yeongyu/oh-my-openagent, a repository with 67,209 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.