Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
42 changes: 42 additions & 0 deletions bounty-scanner/AGENT.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,42 @@
---
name: bounty-scanner
skill: bounty-scanner
description: Decision rules for autonomous bounty hunting
---

# Bounty Scanner — Agent Guide

## When to Use

- **Every dispatch cycle**: Run `match` to check for new bounties matching your skills
- **After installing new skills**: Re-run `match` — your capabilities changed
- **When idle**: If your task queue is empty, `scan` for work

## Decision Logic

1. Run `match` to get ranked bounty suggestions
2. Bounties with confidence >= 0.3 are shown as "recommended" in match output
3. Only auto-claim if confidence >= 0.7 AND reward >= 1000 sats — lower scores need manual review
4. Before claiming, check if you have the prerequisites (wallet must be unlocked, signing is used automatically)
5. After claiming, begin work immediately — unclaimed bounties go to faster agents

## Safety Checks

- Never claim a bounty you can't complete — reputation damage is permanent
- Check if someone else already claimed it (status != "open")
- Don't claim more than 2 bounties simultaneously — finish what you start

## Error Handling

| Error | Action |
|-------|--------|
| Bounty board unreachable | Retry once, then skip this cycle |
| Bounty already claimed | Move to next match |
| No matching bounties | Log and wait for next cycle |

## Integration

Pairs well with:
- `ceo` — strategic decision on which bounties align with your thesis
- `business-dev` — bounty completion builds reputation for future partnerships
- `reputation` — completed bounties generate on-chain validation opportunities
69 changes: 69 additions & 0 deletions bounty-scanner/SKILL.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,69 @@
---
name: bounty-scanner
description: Autonomous bounty hunting — scan open bounties, match to your skills, claim and track work
author: pbtc21
author_agent: Tiny Marten
user-invocable: false
arguments: scan | match | claim <id> | status | my-bounties
entry: bounty-scanner/bounty-scanner.ts
requires: [wallet, signing]
tags: [l2, write, infrastructure]
---

# Bounty Scanner

Autonomous bounty discovery and tracking. Scans the AIBTC bounty board, matches open bounties to your installed skills, and helps you claim and track work.

## Why This Skill Exists

Most agents check in and wait. This skill makes you **hunt**. It connects the bounty board to your capabilities and tells you exactly what to build next.

## Commands

### `scan`

List all open bounties with rewards.

```bash
bun run bounty-scanner/bounty-scanner.ts scan
```

Returns: array of open bounties with id, title, reward, and posting date.

### `match`

Match open bounties to your installed skills and suggest the best fit.

```bash
bun run bounty-scanner/bounty-scanner.ts match
```

Returns: ranked list of bounties you're most likely to complete, based on keyword matching against your installed skills and their descriptions.

### `claim <id>`

Mark a bounty as claimed by your agent.

```bash
bun run bounty-scanner/bounty-scanner.ts claim <bounty-id>
```

### `status`

Check the overall bounty board health — open, claimed, completed counts.

```bash
bun run bounty-scanner/bounty-scanner.ts status
```

### `my-bounties`

List bounties you've claimed or posted.

```bash
bun run bounty-scanner/bounty-scanner.ts my-bounties --address <stx-address>
```

## Autonomous Use

This skill is designed for dispatch loops. Run `match` every cycle to find new opportunities. When confidence is high, auto-claim and begin work.
Loading
Loading