Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Pitch

ASecurity

Pitch — 30-Second Pitch Generator. Distills any idea, product, or project into a tight 30-second SPOKEN pitch — hook, problem, solution, proof, ask — on a hard 75-85 word budget, plus a one-line version. Pulls the five raw materials (who it is for, the pain, the one mechanism, the strongest real proof, the specific ask) from whatever exists — idea text, a repo, a page, the conversation — lists honest inferences [ASSUMED] in a separate ASSUMPTIONS section (beat text stays clean speakable words...

2 stars
0 votes
0 copies
0 views
Added 10/2/2026
ai-agentsgoreact

Works with

cli

Security Analysis

A100/100

Scanned 10/2/2026

$npx -y skills add agentik-os/claude-code-skills --skill pitch --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Pitch?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Pitch
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/agentik-os-pitch/badge)](https://www.skillsdirectory.com/skills/agentik-os-pitch)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
Files
SKILL.md
---
name: pitch
description: "Pitch — 30-Second Pitch Generator. Distills any idea, product, or project into a tight 30-second SPOKEN pitch — hook, problem, solution, proof, ask — on a hard 75-85 word budget, plus a one-line version. Pulls the five raw materials (who it is for, the pain, the one mechanism, the strongest real proof, the specific ask) from whatever exists — idea text, a repo, a page, the conversation — lists honest inferences [ASSUMED] in a separate ASSUMPTIONS section (beat text stays clean speakable words), and never invents numbers or traction: no real proof reads 'none yet'. Use when the user says '/pitch', '/omg-pitch', 'elevator pitch', '30-second pitch', 'pitch this', 'pitch my idea/startup/product', 'how do I explain this in 30 seconds', or in French 'pitch 30 secondes', 'pitche', 'pitche-moi ça', 'présente mon projet en 30 secondes'. NOT for landing pages or long-form marketing copy (that is the marketing suite, e.g. /omg-ad-creative) — /pitch is the spoken 30-second distillation."
---

# Pitch — 30-Second Pitch Generator

Thirty seconds is what you actually get — the elevator, the demo-day mic, the "so what do you do?" at dinner. Pitch compresses whatever exists — a one-line idea, a repo, a landing page, this conversation — into 75-85 words that survive being said out loud: one hook, one problem, one mechanism, one proof, one ask — plus a one-line version for when you get five seconds instead. It is honest by construction: weak material yields a flagged pitch and a note on what to go get, never a polished lie (L2 — researcher, not sycophant). Not for landing pages or long-form marketing copy — that is the marketing suite (/omg-ad-creative); this is the spoken distillation.

---

## Invocation

```
/pitch <idea text | repo path | URL>                        # pitch that
/pitch                                                      # pitch what this conversation is about
/pitch ... --audience [investor|client|partner|recruiter]   # tune proof + ask (protocol step 4)
/omg-pitch ...                                              # same skill, OmegaOS-prefixed
```

No argument and nothing pitchable in context → ask one line ("Pitch what?"). That is the ONLY question this skill ever asks — given any material at all: extract, infer, tag, deliver, zero clarifying questions.

---

## Protocol

### 1. Extract the five raw materials

From whatever the user gave — idea text, a repo (read the README and the core code), a page (fetch and read it), or this conversation — extract:

- **WHO** it is for — one named buyer or user. "Everyone" is not an answer; pick the person who pays.
- **PAIN** — the problem as the WHO feels it, in their words, costed when the source allows.
- **MECHANISM** — how the solution works: ONE mechanism, not a category. "Watches your Stripe webhooks and retries failed charges" beats "an AI billing platform".
- **PROOF** — the single strongest TRUE item the source actually shows: a number, traction, a credential, a verifiably working demo. One. PROOF is never inferred — `[ASSUMED]` is banned outright on numbers, users, revenue, and logos; nothing verifiable in the source → PROOF is `none yet` (the Output Format shows the form).
- **ASK** — what you want from the listener: money, a meeting, a signup. Specific or it does not count. An ask stated in the source is used as-is — it is sourced material, never tagged `[ASSUMED]`. Only when no ask is derivable from the source and no `--audience` is given → default to the demo-day ask (step 4); never ask the user to choose.

Missing WHO, PAIN, MECHANISM, or ASK → infer the strongest HONEST candidate and list the inference solely under ASSUMPTIONS, keyed by material name — never inside beat text. An assumed mechanism is inference; a fabricated metric is a lie.

### 2. Draft to the word budget

75-85 words total — about 30 seconds spoken. The 75-85 total is BINDING; per-beat budgets are targets inside it:

| Beat | Words | Job |
|------|-------|-----|
| Hook | 10-12 | A tension or a number — never "Hi, I am..." |
| Problem | 16-22 | The pain, inside the WHO's world |
| Solution | 22-28 | The one mechanism |
| Proof | 12-15 | One number beats three adjectives |
| Ask | 8-12 | Specific and dated |

PROOF `none yet` → the beat counts zero words; its 12-15 go to PROBLEM and SOLUTION. Count the words. Over 85: cut adjectives first, then whole clauses. Under 75: PROBLEM or PROOF is underfed — expand the costed pain or the proof specifics; never pad with adjectives. Past 90 the draft has failed — recut, do not rationalize.

### 3. Spoken test

Rewrite until the draft survives being read aloud:

- One idea per sentence — a sentence that needs a breath in the middle gets split.
- Kill any word that trips the tongue; if you stumble on it, the speaker will too.
- Zero jargon a stranger would not know — "idempotent retry orchestration" dies here.

### 4. Audience tuning

If the user names an audience, retune PROOF and ASK to it — the other beats barely move:

- **investor** — proof: the growth or market number; ask: the raise and what it buys, dated.
- **client** — proof: the outcome for someone like them; ask: a pilot or demo, dated.
- **partner** — proof: the asset you bring them; ask: the specific integration or intro.
- **recruiter** — proof: the shipped result; ask: the role or the conversation you want.

No audience named → deliver immediately. If the source states an ask, keep it as-is — it is sourced, appears nowhere under ASSUMPTIONS, and the ASSUMPTIONS terminator ("none, all five materials came from the source") stays reachable on a bare `/pitch` run. Only when no ask is derivable from the source does the demo-day default fire: the ASK beat becomes a specific, dated meeting or demo ("give me 15 minutes Thursday to show you the demo"), listed as `[ASSUMED] ASK` under ASSUMPTIONS with the standing retune line shown in the Output Format. Never ask which audience first.

### 5. Deliver

Emit the Output Format block below: the labeled 30s pitch with total word count and approximate seconds, the one-liner ("X helps Y do Z without W" — or sharper when the material allows), delivery notes (where to pause, the one number or phrase to land, plus the mandatory go-get when PROOF holds no true number), and every inference under ASSUMPTIONS. If the material is weak — no real proof, fuzzy WHO — say so next to the pitch and name what to go get. The pitch is only as strong as its weakest material; flattering the user about it helps no one (L2).

---

## Output Format

Emit exactly this block. Beat text is clean speakable words only — `[ASSUMED]` never appears inside a beat or the one-liner; every inference is listed solely under ASSUMPTIONS, keyed by material name (WHO/PAIN/MECHANISM/PROOF/ASK).

```
PITCH (30s) · [total] words · ~[n]s spoken
──────────────────────────────────────────
HOOK     ([n]w)  [text]
PROBLEM  ([n]w)  [text]
SOLUTION ([n]w)  [text]
PROOF    ([n]w)  [text]            <- no true proof in the source: this line reads exactly "PROOF (0w) none yet"
ASK      ([n]w)  [text]

ONE-LINER
[X helps Y do Z without W — or sharper]

DELIVERY NOTES
- Pause: [where, and why it works there]
- Land: [the one number or phrase to hit hardest]
- Go get: [the one specific number that would fill the PROOF beat]   <- mandatory whenever PROOF holds no true number; omit otherwise

ASSUMPTIONS
- [ASSUMED] [material]: [what was inferred, and from what]
- [ASSUMED] ASK: demo-day default (rerun with --audience investor|client|partner|recruiter to retune PROOF+ASK)   <- mandatory whenever the ask was defaulted (no ask in the source and no --audience)
(or: none, all five materials came from the source)
```

Proof-present and proof-absent runs emit the same structure. The PROOF-absent form changes exactly two things — the PROOF line reads `none yet` (words reallocated per step 2) and the `Go get:` note becomes mandatory. It never fills the beat with an assumed number, never drops the beat, never shrinks the block.

---

## Rules & Anti-patterns

**Rules**

- Honesty over flattery (L2). No real number yet → PROOF carries the strongest verifiable status the source actually shows ("the demo works end-to-end", "a waitlist of 40") or reads exactly `none yet` — and the `Go get:` note names the number to earn. Never a polished stand-in.
- `[ASSUMED]` marks inference on WHO/PAIN/MECHANISM/ASK only, and only under ASSUMPTIONS. Numbers, users, revenue, logos: real or absent — never assumed.
- One mechanism. If it cannot be said in one clause, the pitch is not ready — tell the user that instead of shipping mush.
- The word budget is hard. Negotiating it is answering a different question.
- R-NODASH: emitted pitch text (beats, one-liner, notes) uses human punctuation only: comma, period, colon, parentheses. Strip every em or en dash before delivering.

**Anti-patterns — each forces a rewrite**

- Buzzword salad: "revolutionary AI-powered platform" says nothing. Name the mechanism.
- Listing 3 features instead of one mechanism. A list is a catalog, not a pitch.
- Invented or unverifiable numbers. One fake metric poisons every true one.
- Exceeding 90 words. That is a memo, not a pitch.
- Burying or omitting the ask. A pitch without an ask is an anecdote; the ask is the point.

---

## Pairs With

- **/omg-social-content** — hand the finished pitch over: the same five materials become organic posts, threads, and hooks.
- **/omg-cold-email** — the pitch's WHO, PAIN, and ASK are the spine of an outbound sequence; reuse them verbatim.
- **/10x** — when the pitch feels small, the problem is upstream: reset the ambition first, then pitch the 10x version.
- **/ghost** — rehearse it: have the pitch's exact WHO react to the finished 30 seconds before a real one does.

Attribution

agentik-osagentik-os
View sourceSee grades on GitHubMore from agentik-os →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

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', ...

698431 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →