Prompt Pivot Pitch — Agent Context Bundle
================================================================================
spec: agent-context/v1
canonical_url: https://prompt-pivot-pitch.com/agent-context.txt
generated: 2026-06-03T03:39:36.448Z
encoding: UTF-8, LF
content_digest: sha256:e4fc8b413ab5
approx_size: 25.0 KB / ~6394 tokens
================================================================================

WHAT IS THIS
A single, self-contained context bundle for AI agents and LLMs. Paste it
into any chat LLM (no tools needed) or fetch this one URL. Every section
lists an authoritative SOURCE URL you can re-fetch for the live version.

BROWSER-SAFE ALTERNATE
If your browser tool blocks opening this raw .txt, navigate to the rendered
HTML copy instead (identical content): https://prompt-pivot-pitch.com/agent-context.html

HOW TO USE
1. Read SECTION orientation for what the event is and who it serves.
2. Follow SECTION task-rules to shape source material into a problem card.
3. Validate output against SECTION public-card-json-schema.
4. Use SECTION api-summary to read or submit problems.
5. Use SECTION externa-bootstrap to provision a persistent local agent.

SAFETY (non-negotiable)
- Never invent validation, revenue, votes, pledges, pilots, or customer names.
- Never expose raw pledge totals publicly; use pledge/vote signals only.
- Keep private submitter details out of public cards.
- Markdown is canonical for agents; JSON Schema is the contract.

================================================================================
SECTION: orientation
SOURCE: https://prompt-pivot-pitch.com/llms-full.txt
================================================================================

# Prompt Pivot Pitch: Full Content Reference

> Version: 2026-05-14 | Domain: prompt-pivot-pitch.com

---

## Overview

Prompt Pivot Pitch (PP&P) is a Utah AI Build Challenge and AI Problem Studio: a three-event summer series held on Fridays at DF Dance Studio in Salt Lake City, Utah. It bridges the gap between people who have real problems and projects, and technical builders who can apply AI to move that work forward.

The format: anyone can submit a specific problem, a $500 pledge is suggested but not required, the top 10 are selected and published, and competitors work with top AI experts to build the best AI solution across three events culminating in a pitch night.

**Tagline:** Real business problems. AI-powered teams. Clear pitches.

---

## Event Series (Summer 2026)

### Event 1: Prompt (June 12, 2026)
- **Date:** Friday, June 12, 2026
- **Time:** 3:30 PM - 6:00 PM MT
- **Venue:** DF Dance Studio, Salt Lake City, UT
- **Format:**
  - 3:30 PM: Doors open, arrival, networking
  - 4:00 PM: Rotate through 10 selected problem tables (People, Product, Process, Profit, Purpose)
  - 4:35 PM: Team formation, mind-mapping, role assignment
  - 4:50 PM: Venture spec & prompt pack kickoff
  - 5:00 PM: Dinner, networking, beginner Latin or Swing social hour
- **Outcome:** Each team has a problem owner, a build team, and a working venture spec.

### Event 2: Pivot (July 11, 2026)
- **Date:** Friday, July 11, 2026
- **Time:** 3:30 PM - 6:00 PM MT
- **Venue:** DF Dance Studio, Salt Lake City, UT
- **Format:** Teams return with demos-in-progress. Mentor tables test the business case. Pitch decks are reviewed and sharpened.
- **Outcome:** Refined demo, validated business case, investor-ready narrative.

### Event 3: Pitch (August 14, 2026)
- **Date:** Friday, August 14, 2026
- **Time:** 3:30 PM - 6:00 PM MT
- **Venue:** DF Dance Studio, Salt Lake City, UT
- **Format:** Final showcase. Teams pitch real solutions to judges, investors, operators, and sponsors. Followed by dinner, networking, and a social dance hour.
- **Outcome:** Winners recognized, investor connections made, alumni community formed.

---

## Problem Submission

### Who should submit a problem?
Business owners, operators, and domain experts in Utah with a real, specific, AI-applicable operational problem. The problem should be something a business or community faces today that AI could plausibly help solve.

### How to submit
Visit `/submit-problem` on the site. Provide:
- A specific problem or project description (not a product idea; a real business pain)
- The strongest "P" category (People, Product, Process, Profit, or Purpose)
- A suggested $500 pledge, if appropriate (signals commitment; money goes toward prizes and event costs)

### Selection process
Problems are published for community upvoting. Selection factors:
1. Pledged support ($)
2. Community upvotes from operators who face the same problem
3. Category distribution across the 5 Ps
4. Organizer review for specificity, AI applicability, and uniqueness

The top 10 problems are announced before the first event.

---

## The 5 P Categories

- **People**: AI applied to hiring, leadership, learning, customer success, community, and team workflows.
- **Product**: AI applied to building new products, improving existing experiences, and turning domain knowledge into useful features.
- **Process**: AI applied to operations, handoffs, automation, quality, service delivery, and repeatable business systems.
- **Profit**: AI applied to revenue, pricing, margins, forecasting, sales, retention, and investor-ready business models.
- **Purpose**: AI applied to mission-driven goals, trust, access, and community or customer value.

---

## Applying to Build (Participants)

### Who should apply?
- **Business-context contributors**: Business owners, domain experts, operators, or problem owners who can explain the stakes.
- **Technical builders**: engineers, data scientists, no-code automation builders, AI tool users, and AI practitioners.

Every competing team needs business context and technical build capacity. Teams are formed at the first event (Prompt).

### Application
Visit `/register` to apply. Indicate which events you can attend and the role you can play.

---

## Executive Track

A separate track for senior operators and executives who want to observe, mentor, or explore AI applicability for their organizations without joining a competing team. Details at `/executives`.

---

## Challenge Framework

### Application tracks (team composition)
1. **Business + Technical**: Core team. Problem owner or domain expert + technical builder.
2. **Domain Expert + Builder**: Operator with deep domain knowledge + AI practitioner.
3. **Product + Engineering**: Product thinker + builder who can ship.

### AI Applicability Studios (the 5 Ps expanded)
Each problem is placed in one of five studios corresponding to where AI has the most leverage in that business.

### Judging rubric
Teams are evaluated on:
- **Problem clarity**: Is the problem real, specific, and well-understood?
- **AI fit**: Does AI genuinely help, or is it a hammer looking for a nail?
- **Demo quality**: Does the solution actually work in a demo?
- **Business viability**: Can this be monetized, scaled, or adopted by the problem owner?
- **Pitch clarity**: Is the narrative crisp, honest, and investor-ready?

### The 6 Ts of Investability
Teams should demonstrate: **Team**, **Technology**, **Traction**, **Total addressable market**, **Terms**, **Timeline**.

### The 5 Vs of Automation Potential
When evaluating process problems: **Volume**, **Variety**, **Velocity**, **Verifiability**, **Value**.

---

## Partners

### Community Partners
- **DF Dance Studio**: Host venue. The space, movement, and social-hour lesson.
  - https://www.dfdancestudio.com/
- **Utah Geek Events**: Utah community events channel and nonprofit pathway for sponsorship routing.
  - https://www.utahgeekevents.com/
- **Goldman Sachs 10,000 Small Businesses**: Small-business operator participation pathway.
  - https://10ksbapply.com/
- **AI Utah**: Local AI builders and operators.
  - https://luma.com/867qhp9i?tk=cXTp2z
- **City Weekly**: Utah media reach and local storytelling.
  - https://www.cityweekly.net/

### Sponsorship
Organizations can sponsor to gain visibility with Utah's AI-builder and business-operator community. Details at `/sponsor`.

---

## FAQ

**Q: What is Prompt Pivot Pitch?**
A: A Utah AI Build Challenge where anyone can submit a real problem, and competitors work with AI experts across three summer Fridays.

**Q: Can businesses submit problems now?**
A: Yes. Visit /submit-problem. The organizing team helps clarify the problem, choose the best 5 P category, and understand the suggested pledge and selection process.

**Q: Is attendance free?**
A: Participant attendance is application-based. The goal is a curated room with both business context and technical build capacity.

**Q: Do I need dance experience?**
A: No. After the working session there is dinner, networking, and a beginner Latin or Swing lesson. Shoes you can move in are enough; partners are optional.

**Q: Do competing teams need both business and technical roles?**
A: Yes. Teams need business context from a problem owner, operator, or domain expert, plus technical build capacity from AI practitioners, engineers, data people, or automation builders.

**Q: Can I attend only one night?**
A: Apply for the dates you can attend. Competing teams get the most value from the full arc: Prompt forms the team, Pivot sharpens the demo and deck, and Pitch is the final showcase.

**Q: Who leads the social hour?**
A: DF Dance Studio leads the social hour, music, and beginner-friendly lesson.

**Q: How do accessibility and dietary needs work?**
A: Indicate in the registration form. The team batch-plans food and seating around those answers.

**Q: Will my profile be public?**
A: Only if you opt into networking visibility.

---

## Site Map

| Page | URL | Description |
|---|---|---|
| Home | https://prompt-pivot-pitch.com/ | Overview, hero, pillars, partners |
| Events | https://prompt-pivot-pitch.com/events | Event dates, schedule, format |
| Challenge | https://prompt-pivot-pitch.com/challenge | Tracks, rubric, 5 Ps, 6 Ts |
| Submit Problem | https://prompt-pivot-pitch.com/submit-problem | Business problem submission form |
| Register | https://prompt-pivot-pitch.com/register | Builder/founder application |
| Schedule | https://prompt-pivot-pitch.com/schedule | Detailed event program |
| Partners | https://prompt-pivot-pitch.com/partners | Community partner directory |
| Sponsor | https://prompt-pivot-pitch.com/sponsor | Sponsorship information |
| Executive Track | https://prompt-pivot-pitch.com/executives | Senior operator pathway |
| About | https://prompt-pivot-pitch.com/about | Organizer background |
| FAQ | https://prompt-pivot-pitch.com/faq | Frequently asked questions |

---

## Contact & Social

- Email: baguleyllc@gmail.com
- Hashtag: #promptpivotpitch
- Sitemap: https://prompt-pivot-pitch.com/sitemap.xml

---

## Prompt Pivot Pitch Agent Context System

Markdown is the canonical source for agents. HTML is the rendered human experience. JSON Schema is the contract.

Agents should help users turn source material into:

1. A public voting card.
2. A private readiness brief.
3. A structured JSON submission payload.
4. Share copy.
5. One unresolved question.

Use "source material" or "problem context" for user-provided input. Users may provide a short description, well-formed problem, full PRD, technical spec, customer interview transcript, or implementation memo.

### Agent lookup order

0. `https://prompt-pivot-pitch.com/agent-context.txt` for a single-fetch bundle of everything below (orientation, task rules, schema, and API summary). Best for chat-only LLMs with no tools.
1. `https://prompt-pivot-pitch.com/llms.txt` for orientation.
2. `https://prompt-pivot-pitch.com/llms-full.txt` for complete context.
3. `https://prompt-pivot-pitch.com/problem-shaper.md` for task rules.
4. `https://prompt-pivot-pitch.com/problem-card-schema.json` for public-card validation.
5. `https://prompt-pivot-pitch.com/openapi/problem-api-v1.json` for endpoint contracts.
6. Swagger, Redoc, or Fern-generated SDKs/CLI when available.

### Core problem API

- `GET /api/problem/v1/schema`
- `GET /api/problem/v1/criteria`
- `POST /api/problem/v1/shape-preview`
- `POST /api/problem/v1/submit`
- `GET /api/problem/v1/submissions/{id}/status`
- `GET /api/problem/v1/submissions/{id}/public-card`
- `GET /api/problem/v1/submissions/{id}/share-copy`
- `GET /api/problem/v1/submissions/{id}/pledge-signal`
- `GET /api/agent/v1/bootstrap?submissionId={id}&format=markdown`
- `GET /api/agent/v1/bootstrap/preview?submissionId={id}`
- `GET /api/agent/v1/bootstrap/files?submissionId={id}`
- `GET /api/agent/v1/llmcopy?submissionId={id}`
- `GET /api/agent/v1/discover`
- `GET /api/agent/mcp/manifest`

### Safe pledge and vote signals

Do not expose raw pledge totals publicly. Public agents should use this shape:

```json
{
  "submissionId": "problem_123",
  "pledgeSignal": "none | interest | pilot_interest | sponsor_interest | private",
  "voteSignal": "none | early_interest | active_interest | strong_interest",
  "publicSummary": "This problem has received private interest.",
  "privateDetailsAvailable": false
}
```

### Nontechnical local-agent flow

1. Ask the user for source material.
2. Call `POST /api/problem/v1/shape-preview`.
3. Explain the public card first.
4. Keep private readiness notes separate.
5. Offer to submit, revise, generate share copy, or track the submission.

### Technical local-agent flow

1. Inspect OpenAPI and schema.
2. Validate JSON payloads before submission.
3. Use Swagger "try it out", Redoc, Fern SDK/CLI, `curl`, or a local script.
4. Store submission ids in local state.
5. Use status, public-card, share-copy, vote-signal, and pledge-signal endpoints to keep the user informed.

### Externa ExecOS posture

Externa acts as Chief of Staff working directly with the CEO for the user's Prompt Pivot workspace. It can operate at a practical executive level across CTO, CFO, COO, CMO, product, strategy, operations, and related roles.

When a task has an executive domain, Externa should name the role lens, take a current timestamp, and look up current best practices if web or API access is available. If current lookup is unavailable, it should say so and use stable principles only.

### Share-copy safety

Share copy must use only public-card content, official status, public vote signal, public pledge signal, and user-approved facts. It must not include private readiness notes, private submitter context, private customer names, emails, phone numbers, raw pledge totals, or private pledge details.

Do not imply endorsement, selection, funding, partnership, availability, urgency, sponsorship, pilots, revenue, customer demand, or organizer approval unless verified. Ask for human approval before posting, emailing, texting, tagging people, or contacting sponsors/builders.

### Local Externa agent kit

- `https://prompt-pivot-pitch.com/api/agent/v1/bootstrap?format=markdown`
- `https://prompt-pivot-pitch.com/api/agent/v1/bootstrap/preview`
- `https://prompt-pivot-pitch.com/api/agent/v1/bootstrap/files`
- `https://prompt-pivot-pitch.com/prompt-pivot-agent/SKILL.md`
- `https://prompt-pivot-pitch.com/prompt-pivot-agent/onboarding.md`
- `https://prompt-pivot-pitch.com/prompt-pivot-agent/watch-submission.md`

The primary Externa workflow is paste-to-provision. The user copies one bootstrap prompt into ChatGPT, Claude, Cursor, Gemini, or a local agent. That assistant creates or shows the workspace files wherever the user chooses. Direct file downloads are an advanced fallback.

================================================================================
SECTION: task-rules
SOURCE: https://prompt-pivot-pitch.com/problem-shaper.md
================================================================================

# Prompt Pivot Pitch Problem Shaper

Prompt Pivot Pitch is a Utah AI Problem Studio where operators, founders, hackers, and community partners turn real problems into buildable AI projects. Businesses submit problems, stalled projects, community needs, or product opportunities. Founder plus hacker teams shape, build, pivot, and pitch AI-enabled solutions across Summer 2026 events hosted at DF Dance Studio in Salt Lake City.

Markdown is the canonical source for agents. HTML is the rendered human experience. JSON Schema is the contract.

## Your Job

Convert the user's source material into:

1. A public voting card.
2. A private readiness brief.
3. A structured JSON payload.
4. Share copy.
5. One unresolved question.

Use "source material" or "problem context"; do not describe user input as unfinished or informal. The user may paste a mature PRD, spec, memo, transcript, or a short description.

## Clarifying Questions

Ask at most one clarifying question only when a critical input is missing:

- who feels it
- today's state
- what AI changes
- current work stage
- data or integration requirements
- evidence of demand

If the source material is usable, do not ask a question before drafting.

## Public Card Headers

Use these exact public-card headers:

**Title.**
**5 P category.**
**Value proposition.**
**Who feels it.**
**Today's state.**
**Market signal.**
**What AI shifts.**
**Build readiness flags.**
**Demand signals.**
**Better path, if applicable.**

End the public card with one short unresolved question.

## 5 P Categories

- **People**: hiring, leadership, learning, customer success, community, and team workflows.
- **Product**: new products, improved experiences, and domain knowledge turned into useful features.
- **Process**: operations, handoffs, automation, quality, service delivery, and repeatable systems.
- **Profit**: revenue, pricing, margins, forecasting, sales, retention, and investable business models.
- **Purpose**: mission-driven goals, trust, access, public benefit, and community or customer value.

## Readiness Checks

Evaluate:

- PDLC stage and current work state.
- SDLC feasibility and first build slice.
- Investability and demand evidence.
- Data availability and data quality.
- Privacy, consent, and regulated-data constraints.
- Human review checkpoints.
- Graceful failure modes and confidence boundaries.

## Rules

- Keep public cards under 450 words.
- Use plain language.
- Do not use hype adjectives.
- Do not invent numbers, pledges, votes, customer names, revenue, pilots, or validation.
- Replace private names, emails, phone numbers, and private customer details with generic placeholders.
- Separate public-card content from private readiness notes.
- If the real issue is people, process, or organizational rather than technical, say so.
- Mention privacy, consent, data quality, human review, and graceful failure when relevant.
- Never expose raw pledge totals publicly.

## Browser-Safe HTML Alternate

If your browser tool blocks opening this raw `.md` file as an "unsafe direct URL", navigate to the identical HTML copy instead:
`https://prompt-pivot-pitch.com/problem-shaper.html`

## API Use

Base URL: `https://prompt-pivot-pitch.com`

Recommended lookup order for agents:

1. `/llms.txt` for orientation.
2. `/llms-full.txt` for full event and API context.
3. `/problem-shaper.md` (or `/problem-shaper.html`) for task rules.
4. `/problem-card-schema.json` for public-card validation.
5. `/openapi/problem-api-v1.json` for endpoint contracts.
6. Swagger, Redoc, or Fern-generated docs/SDKs when available.

### Problem board (for agents that cannot execute JavaScript)

The public problem board at `/submit-problem` is a JavaScript-rendered React app. Agents should fetch board data from the API directly instead of rendering the page:

- `GET /api/problems` — list all public problems (unauthenticated, no private fields)
- `GET /api/problems/{id}` — get a single public problem by id

### Core endpoints:

- `GET /api/problems`
- `GET /api/problems/{id}`
- `GET /api/problem/v1/schema`
- `GET /api/problem/v1/criteria`
- `POST /api/problem/v1/shape-preview`
- `POST /api/problem/v1/submit`
- `GET /api/problem/v1/submissions/{id}/status`
- `GET /api/problem/v1/submissions/{id}/public-card`
- `GET /api/problem/v1/submissions/{id}/share-copy`
- `GET /api/problem/v1/submissions/{id}/pledge-signal`
- `GET /api/agent/v1/bootstrap?submissionId={id}&format=markdown`
- `GET /api/agent/v1/bootstrap/preview?submissionId={id}`
- `GET /api/agent/v1/bootstrap/files?submissionId={id}`
- `GET /api/agent/v1/llmcopy?submissionId={id}`
- `GET /api/agent/v1/discover`
- `GET /api/agent/mcp/manifest`

Public pledge signal shape:

```json
{
  "submissionId": "problem_123",
  "pledgeSignal": "none | interest | pilot_interest | sponsor_interest | private",
  "voteSignal": "none | early_interest | active_interest | strong_interest",
  "publicSummary": "This problem has received private interest.",
  "privateDetailsAvailable": false
}
```

Do not request or reveal raw pledge totals unless an authenticated submitter-only API explicitly returns them.

## Example Shape Preview Request

```bash
curl -X POST https://prompt-pivot-pitch.com/api/problem/v1/shape-preview \
  -H "Content-Type: application/json" \
  -d '{
    "source": "agent_api",
    "sourceMaterial": "Paste the problem context, PRD, spec, transcript, or notes here.",
    "title": "Optional working title"
  }'
```

## Expected Output

The response includes:

- `publicCard`
- `privateReadiness`
- `submissionPayload`
- `shareCopy`
- `unresolvedQuestion`
- `publicSafety`

## Externa Bootstrap

Use Externa Bootstrap when a user wants their own AI assistant to keep helping after shaping or submission. The endpoint returns one pasteable provisioning prompt:

```bash
curl "https://prompt-pivot-pitch.com/api/agent/v1/bootstrap?submissionId=problem_123&format=markdown"
```

The user pastes `paste_blob` into ChatGPT, Claude, Cursor, Gemini, or a local agent. That assistant creates or shows a local workspace with `BOOTSTRAP.md`, `frontmatter.yaml`, `USER.md`, `IDENTITY.md`, `SOUL.md`, `AGENTS.md`, `TOOLS.md`, `HEARTBEAT.md`, `state/`, and `skills/`.

================================================================================
SECTION: public-card-json-schema
SOURCE: https://prompt-pivot-pitch.com/problem-card-schema.json
================================================================================

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://prompt-pivot-pitch.com/problem-card-schema.json",
  "title": "Prompt Pivot Pitch Public Problem Card",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "title",
    "fivePPrimary",
    "valueProposition",
    "whoFeelsIt",
    "todaysState",
    "marketSignal",
    "whatAIShifts",
    "buildReadinessFlags",
    "demandSignals",
    "unresolvedQuestion"
  ],
  "properties": {
    "title": { "type": "string", "minLength": 3, "maxLength": 90 },
    "fivePPrimary": { "$ref": "#/$defs/fiveP" },
    "fivePSecondary": { "$ref": "#/$defs/fiveP" },
    "valueProposition": { "type": "string", "minLength": 10, "maxLength": 500 },
    "whoFeelsIt": { "type": "string", "minLength": 10, "maxLength": 500 },
    "todaysState": { "type": "string", "minLength": 10, "maxLength": 900 },
    "marketSignal": { "type": "string", "minLength": 10, "maxLength": 700 },
    "whatAIShifts": { "type": "string", "minLength": 10, "maxLength": 800 },
    "buildReadinessFlags": { "type": "string", "minLength": 10, "maxLength": 900 },
    "demandSignals": { "type": "string", "minLength": 10, "maxLength": 600 },
    "betterPathIfApplicable": { "type": "string", "maxLength": 500 },
    "unresolvedQuestion": { "type": "string", "minLength": 8, "maxLength": 220 }
  },
  "$defs": {
    "fiveP": {
      "type": "string",
      "enum": ["People", "Product", "Process", "Profit", "Purpose"]
    }
  },
  "x-prompt-pivot-constraints": [
    "Keep public cards under 450 words.",
    "Do not include private submitter emails, phone numbers, or private customer names.",
    "Do not invent validation, revenue, votes, pledges, pilots, or customer names.",
    "Do not expose raw pledge totals publicly."
  ]
}

================================================================================
SECTION: api-summary
SOURCE: https://prompt-pivot-pitch.com/openapi/problem-api-v1.json
================================================================================

Prompt Pivot Pitch Problem API (v1.0.0)
API-first contract for problem shaping, public cards, safe share copy, and agent discovery.
Server: https://prompt-pivot-pitch.com/api

Endpoints:
- GET /problem/v1/schema [ppp_get_problem_schema] — Get public problem-card schema
- GET /problem/v1/criteria [ppp_get_criteria] — Get problem-shaping criteria
- POST /problem/v1/shape-preview [ppp_shape_problem] — Shape source material into public/private outputs
- POST /problem/v1/submit [ppp_submit_problem] — Submit a problem
- GET /problem/v1/submissions/{id}/status [ppp_get_submission_status] — Get public submission status
- GET /problem/v1/submissions/{id}/public-card [ppp_get_public_card] — Get public card
- GET /problem/v1/submissions/{id}/share-copy [ppp_get_share_copy] — Get public-safe share copy
- GET /problem/v1/submissions/{id}/pledge-signal [ppp_get_pledge_signal] — Get public-safe vote and pledge signals
- GET /problems [ppp_list_public_problems] — List public problem submissions
- GET /problems/{id} [ppp_get_public_problem] — Get one public problem submission
- PATCH /problems/{id} [ppp_update_owned_problem] — Update an owned problem using a private manage token
- POST /problems/{id}/archive-request [ppp_request_problem_archive] — Request organizer-approved archive for an owned problem
- GET /agent/v1/discover [ppp_agent_discover] — Discover agent-readable docs and tools
- GET /agent/v1/bootstrap [ppp_get_externa_bootstrap] — Get paste-to-provision Externa Bootstrap prompt
- GET /agent/v1/bootstrap/files [ppp_get_externa_bootstrap_files] — Get Externa Bootstrap file templates
- GET /agent/v1/bootstrap/preview [ppp_get_externa_bootstrap_preview] — Preview Externa Bootstrap outcome
- GET /agent/v1/llmcopy [ppp_get_llmcopy] — Get LLMCopy paste blob and docs
- GET /agent/mcp/manifest [ppp_mcp_manifest] — Get MCP-ready manifest preview

Full machine-readable contract: see SOURCE URL above.

================================================================================
SECTION: externa-bootstrap
SOURCE: https://prompt-pivot-pitch.com/api/agent/v1/bootstrap?format=markdown
================================================================================

Provision a user's own AI assistant to keep helping after shaping or
submission. Fetch one paste-to-provision prompt:

  GET https://prompt-pivot-pitch.com/api/agent/v1/bootstrap?format=markdown
  GET https://prompt-pivot-pitch.com/api/agent/v1/bootstrap?submissionId={id}&format=markdown

Paste the returned `paste_blob` into ChatGPT, Claude, Cursor, Gemini, or a
local agent. It creates a local Externa workspace (BOOTSTRAP.md,
frontmatter.yaml, USER.md, IDENTITY.md, SOUL.md, AGENTS.md, TOOLS.md,
HEARTBEAT.md, state/, skills/). No live repo, site, or API is mutated.