# Sf Loop Hunter

> Recurring scanner that watches the StrategyFactory HOT leaderboard and surfaces NEW top-N picks as they appear. Designed to be wrapped by the /loop skill (or any cron) — runs one pass per invocation, persists the previous top-N to disk, and emits a delta. Use when the user asks "keep an eye on SF for new winners / monitor the leaderboard / alert me when a new top edge shows up / find me new bange…

- **Type:** Skill
- **Install:** `agentstack add skill-daviddme-claude-skills-sf-loop-hunter`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [daviddme](https://agentstack.voostack.com/s/daviddme)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [daviddme](https://github.com/daviddme)
- **Source:** https://github.com/daviddme/claude-skills/tree/main/SF-Skills/sf-loop-hunter

## Install

```sh
agentstack add skill-daviddme-claude-skills-sf-loop-hunter
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# sf-loop-hunter

The "always-on edge hunter". One invocation = one scan. State persists at
`~/.claude/skills/sf-skills/hunter-state.json` so consecutive runs can
diff "what's new in the top N".

Pair with the [`/loop` skill](../../README.md#loop-driven-skills) to run
on an interval:

```
/loop 15m python SF-Skills/sf-loop-hunter/hunter.py --top 10
```

## What each run does

1. Run `sf-best-pick` logic against the HOT leaderboard (composite score).
2. Load the previous top-N from disk.
3. Diff: which slugs entered the top-N, which dropped out, which moved.
4. Emit a short report. If nothing changed, prints `no change`.
5. Save the new top-N to disk.

## When to use

- "Watch the SF leaderboard and tell me when something new ranks"
- "Run an edge-hunter in the background during the stream"
- "Alert me when a fresh top strategy appears"

## How to run

```
python sf-loop-hunter/hunter.py [--top N] [--state PATH] [--json]
```

Defaults: top 10, state at `~/.claude/skills/sf-skills/hunter-state.json`.

## Output

```
HUNTER PASS @ 2026-05-24T14:00:00Z
  + NEW:     spicy-trend-v2  score=0.83  PF=2.4 30d=+12.1%
  + RISER:   meanrev-eth     score=0.79 (was 0.71, +0.08)
  - DROPPED: tired-edge-v1   was rank 4 — fell out of top 10
```

## Notes

- State is local JSON, easy to inspect. Delete it to "forget" history.
- This is the discovery side. For "is what I'm running still good?" use
  [`sf-loop-pruner`](../sf-loop-pruner/SKILL.md).

## Source & license

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

- **Author:** [daviddme](https://github.com/daviddme)
- **Source:** [daviddme/claude-skills](https://github.com/daviddme/claude-skills)
- **License:** MIT

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

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-daviddme-claude-skills-sf-loop-hunter
- Seller: https://agentstack.voostack.com/s/daviddme
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.