Amplify

Amplify

Turn a rough prompt into a high-signal one. Detect the domain, inject the standards that domain demands, score it against a rubric — then offer to run it.

Dispatcher and reviewer — Opus 4.8 (thinking-tier). Searcher/Writer — Sonnet 4.6 (worker-tier).

Amplify does one thing well: it rewrites the prompt you give it into the single strongest version, then hands it off. It does not write code itself — it produces a prompt other skills (or you) execute.

Per-Step Agent Map (DOCTRINE rule 12)

Every substantive step dispatches at least one Agent. Atomic steps (DOCTRINE 12.2.8) are a single Worker → Reviewer pair with no independent angles to fan out.

| Step | Status | Worker tier | Thinking tier | Notes | |---|---|---|---|---| | 1 — Read intent | Atomic (12.2.8) | Searcher (Sonnet) — domain signals + project rules | Analyst (Opus) — triage + gap analysis + persona selection | Single Worker → Reviewer; reads CLAUDE.md, AGENTS.md, .hyperflow/memory/* | | 2 — Amplify | Atomic (12.2.8) | Writer (Sonnet) — draft the enhanced prompt | Reviewer (Opus) — rubric score, one targeted revision | Writer drafts → Opus scores against the 8-dim rubric → revise once if any dim Run this amplified prompt now? > - Send to spec (recommended) — start the chain design-first; spec brainstorms + designs with the amplified prompt, then auto-chains to scope → dispatch > - Send to scope — the approach is clear, go straight to decomposition (/hyperflow:scope) > - Send to dispatch — a task file already exists (/hyperflow:dispatch) > - Copy only — keep the prompt, run nothing

This is a multi-option gate (4 options) → it carries the (Recommended) marker on Send to spec (DOCTRINE rule 8 — named-workflow choice).

On a Send to … selection, invoke the chosen skill via the Skill tool, passing the amplified prompt as the argument. Spec is the default because a freshly-amplified prompt is a design starting point — spec's brainstorming + section-by-section design is exactly what it feeds. On Copy only, stop — the prompt is already printed for the user to take.

Codex fallback: if the host does not expose the AskUserQuestion popup UI, print the same gate as a Hyperflow Question chat block with the four numbered choices and wait for the user's reply. Do not auto-select Send to spec.

Iron Rules

• Amplify never writes code. It produces a prompt; downstream skills execute it. If the user wanted code, the handoff gate routes there.
• Project rules win on conflict. A rule in CLAUDE.md / AGENTS.md / .hyperflow/memory/ overrides the generic persona standard — it is the user's explicit instruction.
• Economy is mandatory (rubric dim 8). Never inflate a trivial prompt. The rubric enhances to the task's level, not beyond.
• Failure recovery (rule 14). Worker/Reviewer errors and the NEEDS_REVISION cadence follow [../hyperflow/failure-recovery.md](../hyperflow/failure-recovery.md).
• No AI attribution in the amplified prompt or rationale — describe the work, never the author.

Doctrine

Shared rules in [../hyperflow/DOCTRINE.md](../hyperflow/DOCTRINE.md). Personas in [../hyperflow/personas-A.md](../hyperflow/personas-A.md) + [personas-B.md](../hyperflow/personas-B.md). Rubric in [references/prompt-rubric.md](references/prompt-rubric.md).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: jeremylongshore
• Source: jeremylongshore/claude-code-plugins-plus-skills
• License: MIT
• Homepage: https://tonsofskills.com

Install and usage instructions live in the source repository linked above.