79 lines
3 KiB
Markdown
79 lines
3 KiB
Markdown
[//]: # (SPDX-License-Identifier: MIT)
|
|
[//]: # (Copyright (c) 2026 Rug Munch Media LLC)
|
|
# AGENTS.md — PryScraper
|
|
|
|
> AI agent contract. Read this before touching anything in this repo.
|
|
|
|
## Status
|
|
canonical · owner=crmuncher · last_updated=2026-07-06 · phase_0_hardening_complete
|
|
|
|
## What This Repo Is
|
|
Multi-source crypto intelligence crawler (pryscraper). Scraper, extractor, parser, automator, job queue. CLI + MCP server + SDK. License: Dual -- MIT (core) + BSL 1.1 (stealth/anti-detection subset). See LICENSE and LICENSE-BSL-STEALTH.
|
|
|
|
## Type
|
|
`backend` · language=Python 3.12 + FastAPI
|
|
|
|
## Where It Lives
|
|
- Source: `https://git.rugmunch.io/RugMunchMedia/pryscraper`
|
|
- Deployed: Docker on Talos (/srv/pry/)
|
|
- Port: 8005
|
|
|
|
## Built With
|
|
|
|
## Components
|
|
- `/srv/pry/*.py`: 83 Python modules (api.py, extractor.py, parser.py, etc.)
|
|
- `browser-extension/`: Chrome/Firefox extension
|
|
- `routers/`: FastAPI routers split from `api.py` (scraping, templates, health, auth, captcha, …)
|
|
- `pry_mcp/`: MCP sub-package (split from `mcp_production.py` 2026-07-06)
|
|
- `pry_x402/`: x402 constants + enums (split from `x402.py` 2026-07-06)
|
|
- `pry_<name>.py`: 19 root modules renamed to remove `routers/` collision (2026-07-06)
|
|
|
|
## Dependencies
|
|
- postgresql (job storage)
|
|
- redis (cache)
|
|
- flaresolverr (Cloudflare bypass, on Hydra)
|
|
- playwright (browser automation)
|
|
- cloudscraper (Cloudflare bypass tier)
|
|
- aiohttp-socks + aiohttp (SOCKS5 proxy support, async HTTP)
|
|
- pyjwt (JWT auth signing)
|
|
- apify (Apify actor SDK)
|
|
- playwright-stealth (anti-detection)
|
|
|
|
## 🚨 CRITICAL RULES
|
|
|
|
1. **No nested repos.** Don't commit other complete project trees inside this one.
|
|
2. **No secrets.** Never commit `.secrets/`, `.env`, `*.pem`, `*.key`, `*.crt`. Use gopass.
|
|
3. **No data blobs.** Don't commit `*.zip`, `*.parquet`, model weights, sqlite files.
|
|
4. **No broken files.** Shell heredoc accidents must be caught before commit.
|
|
5. **No duplicate scaffolds.** If a `backend/` or `frontend/` subdir exists at root, it's either THE app or bloat.
|
|
6. **Update STATUS.md before committing.** Builders must know where we are.
|
|
7. **Read DECISIONS.md before architectural changes.** Don't repeat rejected designs.
|
|
8. **Respect dual licensing.** When contributing, the core/ files are MIT (any PR welcome). The stealth/ files (listed in LICENSE-BSL-STEALTH) are BSL 1.1 -- only Rug Munch Media LLC can modify them. PRs touching BSL files will be rejected.
|
|
|
|
## Commands
|
|
```bash
|
|
make install # install deps
|
|
make dev # dev server
|
|
make test # tests
|
|
make lint # ruff + format check
|
|
make typecheck # mypy strict
|
|
make security # bandit + gitleaks
|
|
make ci # full pipeline
|
|
make commit # conventional commit
|
|
make clean # remove build artifacts
|
|
```
|
|
|
|
## Architecture
|
|
See [ARCHITECTURE.md](ARCHITECTURE.md). Update it when components change.
|
|
|
|
## Plan
|
|
[PLAN.md](PLAN.md) shows current sprint. Update weekly.
|
|
|
|
## Status
|
|
[STATUS.md](STATUS.md) shows where we are RIGHT NOW. Update before each commit.
|
|
|
|
## Owners
|
|
- Primary: crmuncher
|
|
- Backup: TBD
|
|
|
|
## Related Repos
|