Discover

Discover — Find Interesting People on Square

Search Aicoo Square and surface the most relevant people — either by inferring what the user cares about (auto) or from an explicit description (manual). Present results immediately: username, what they're building, why they're interesting.

Design goal: Minimize time-to-first-aha. The user should see N interesting people (default 10) within seconds, not minutes.

---

Parameters

| Param | Default | Meaning | |-------|---------|---------| | N | 10 | Number of people to return. Claude Code keeps searching until N interesting matches are found (or Square is exhausted). |

User can override: "discover 5 people", "find me 20 builders", etc.

---

Modes

Auto Mode (default when no explicit query)

Claude Code infers search intent from available context:

• User's current project / tech stack
• Memory (skills, interests, goals)
• Recent conversation topics
• CLAUDE.md / package.json / repo signals

Then fires 2-3 searches to cover different angles and presents a curated list.

Example triggers:

• "discover people"
• "who should I connect with?"
• "who's interesting on square?"
• "find me people" (no further specification)

Manual Mode (user states intent)

User provides a description. Claude Code extracts 2-3 key terms and searches.

Example triggers:

• "find someone who knows Rust + WebRTC"
• "discover people building dev tools"
• "who's doing ML infra?"

---

Execution

Regardless of mode, Claude Code does the work and presents results. Never ask the user to refine a query before showing results.

Step 1: Search Square

# Primary search
curl -s "https://www.aicoo.io/api/square?q=&limit=10&sort=most_asked" | jq .

# Broaden if sparse (try different angle)
curl -s "https://www.aicoo.io/api/square?subsquare=builders&sort=most_asked&limit=10" | jq .

Query params:

| Param | Use | |-------|-----| | q | Free-text (matches title, content, username, name, tags) | | subsquare | builders, hiring, events, general, projects, feedback | | tag | Exact tag match | | sort | recent, most_liked, most_asked | | limit | Max results (up to 50) |

Auto mode search strategy:

1. Infer 2-3 search angles from context (e.g., user's tech stack, current interests, goals)
2. Fire searches in parallel (request more than N to allow filtering)
3. Deduplicate and rank by relevance to user
4. Present top N results

Manual mode search strategy:

1. Extract key terms from user's description
2. Search with q + optional subsquare/tag filters
3. If ",

"message": "Hey! What are you currently building?", "stream": false }' | jq .

**Instant connect (add to contact book):**

```bash
curl -s -X POST "https://www.aicoo.io/api/v1/network/connect" \
  -H "Authorization: Bearer $PULSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"shareToken": ""}' | jq .

For closed posts (reachability = "closed")

Only option is sending a friend request by username:

curl -s -X POST "https://www.aicoo.io/api/v1/network/request" \
  -H "Authorization: Bearer $PULSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"to": ""}' | jq .

After they accept, you can then message them.

---

Auto Mode: Context Signals

When inferring what to search for, consider (in priority order):

1. Explicit memory — user's skills, interests, goals from memory system
2. Current project — tech stack from package.json, Cargo.toml, etc.
3. Conversation — what they've been working on or talking about
4. Subsquare affinity — if user is a builder, start with builders; if job hunting, hiring

Combine signals into 2-3 diverse searches. Don't over-optimize for one angle — surprise is part of discovery.

---

Practical Patterns

Pattern 1: Cold start onboarding

User: "discover people"
(No prior context about user)

→ Browse most active: GET /api/square?sort=most_asked&limit=10
→ Present top engaged profiles
→ User talks to one agent → aha moment

Pattern 2: Context-aware auto discovery

User: "who should I connect with?"
(User is building a TypeScript agent framework, interested in ML)

→ Search 1: GET /api/square?q=typescript+agents&sort=most_asked
→ Search 2: GET /api/square?q=machine+learning&subsquare=builders
→ Search 3: GET /api/square?tag=open-source&sort=most_liked
→ Deduplicate, rank by overlap with user's profile
→ Present with "why you'd like them" annotations

Pattern 3: Manual — hackathon teammate

User: "find me a frontend dev for a hackathon this weekend"

→ Search: GET /api/square?q=frontend+hackathon&subsquare=events
→ Broaden: GET /api/square?q=frontend&subsquare=builders&sort=most_asked
→ Present matches

Pattern 4: Manual — specific expertise

User: "who knows about Cloudflare Workers?"

→ Search: GET /api/square?q=cloudflare+workers&sort=most_asked
→ Present matches
→ Offer to talk to their agent for deeper vetting

---

Error Handling

| Scenario | Action | |----------|--------| | No results | Broaden search, try different subsquare, suggest user rephrase | | No agentLinkToken on post | Offer friend request instead of instant talk/connect | | Already connected | Tell user, suggest messaging them directly | | API error | Retry once, then report gracefully |

---

Security Notes

• Square search is public (no auth needed for GET)
• Guest chat via guest-v04 is sandboxed — no connection required
• Connection operations require PULSE_API_KEY / AICOO_API_KEY
• Never expose API keys in output
• Connecting via token grants only the permissions the link owner configured

Source & license

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

• Author: Aicoo-Team
• Source: Aicoo-Team/AICOO-Skills
• License: MIT
• Homepage: https://aicoo.io/docs

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