{"id":50602360,"url":"https://github.com/travisbreaks/samaritan","last_synced_at":"2026-06-05T19:33:13.797Z","repository":{"id":345655839,"uuid":"1186785647","full_name":"travisbreaks/samaritan","owner":"travisbreaks","description":"Multi-agent development system with persistent memory, parallel execution, mobile oversight, and behavioral guardrails.","archived":false,"fork":false,"pushed_at":"2026-03-20T03:20:07.000Z","size":15,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-04-04T03:35:49.223Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/travisbreaks.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-20T01:41:05.000Z","updated_at":"2026-03-20T03:20:10.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/travisbreaks/samaritan","commit_stats":null,"previous_names":["travisbreaks/samaritan"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/travisbreaks/samaritan","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/travisbreaks%2Fsamaritan","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/travisbreaks%2Fsamaritan/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/travisbreaks%2Fsamaritan/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/travisbreaks%2Fsamaritan/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/travisbreaks","download_url":"https://codeload.github.com/travisbreaks/samaritan/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/travisbreaks%2Fsamaritan/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33957498,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-05T02:00:06.157Z","response_time":120,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-06-05T19:33:12.989Z","updated_at":"2026-06-05T19:33:13.792Z","avatar_url":"https://github.com/travisbreaks.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# Samaritan\n\nTemplates, guides, and hard-won guardrails for running AI coding agents as development collaborators.\n\nIf you are building with [Claude Code](https://claude.ai/claude-code) and want to skip the part where you learn these lessons by losing data, deploying secrets, or exhausting your CI quota, start here.\n\n## What's Inside\n\n### [PROTOCOL.md](PROTOCOL.md)\nEight operating principles for working with AI agents responsibly. Each one exists because something broke without it.\n\n### [docs/mobile-access.md](docs/mobile-access.md)\nHow to control your agents from your phone. Blink Shell + mosh + Tailscale + tmux. Survives sleep, wifi/cell switching, spotty signal. Full setup guide for Mac and Linux.\n\n### [docs/galaxy-brain.md](docs/galaxy-brain.md)\nHow to write discussion answers that actually get marked on GitHub. Repo selection, question targeting, answer structure.\n\n### [docs/anti-ai-policy-watchlist.md](docs/anti-ai-policy-watchlist.md)\nProjects that ban or restrict AI contributions. Check before submitting.\n\n### [templates/CLAUDE.md.template](templates/CLAUDE.md.template)\nStarter instruction file for Claude Code with guardrails pre-configured.\n\n### [memory/oss-targets.template.md](memory/oss-targets.template.md)\nTracking template for open source contribution work.\n\n## What Goes Wrong\n\nThese are real incidents from production use:\n\n1. **A chained bash command wiped ~1TB.** One semicolon, two commands, wrong target. Rule: one command per call, no chaining.\n2. **Agent deployed client financial data to a public URL.** No concept of data classification. Rule: third-party data stays local or behind auth.\n3. **Context compaction destroyed in-flight work.** Compressed context lost critical state. Rule: checkpoint every 45 minutes.\n4. **Concurrent sessions overwrote each other's git changes.** No cross-session awareness. Rule: each session gets its own branch.\n5. **Agent exhausted monthly CI quota in 18 pushes.** No model of finite resources. Rule: validate locally before pushing.\n\nEvery guardrail traces to an incident. The failures are the documentation.\n\n## Architecture (Reference)\n\nThis is one working configuration. Adapt to what you have.\n\n```\nPhone (Blink Shell + mosh)\n  |\n  Tailscale (encrypted mesh, no port forwarding)\n  |\n  Dev Machine\n  |--- VS Code + Claude Code (primary agent)\n  |--- Parallel subagents (scout, review, research, code)\n  |--- Docker (isolated worker agents, optional)\n  |--- tmux (persistent sessions)\n  |\n  Cloud Server (optional)\n  |--- Watchdog agent (health checks)\n  |--- Scout agent (research, monitoring)\n```\n\nThe minimum viable setup is Claude Code + a CLAUDE.md file + a memory directory. Everything else is optional and additive.\n\n## Getting Started\n\n1. Install [Claude Code](https://claude.ai/claude-code) (VS Code extension or CLI)\n2. Copy [templates/CLAUDE.md.template](templates/CLAUDE.md.template) to your project root as `CLAUDE.md`\n3. Create a `memory/` directory for persistent state\n4. Read [PROTOCOL.md](PROTOCOL.md) before doing anything else\n\nFor mobile access, follow [docs/mobile-access.md](docs/mobile-access.md).\n\n## Philosophy\n\nAI agents are powerful local optimizers operating without global awareness. They will confidently do the wrong thing if the guardrails are not there. The system provides global awareness: rules, memory, oversight. The agents provide local execution: research, code, analysis. The human provides judgment.\n\nThe agent swarm is not autonomous. It is an instrument. The human plays it.\n\n## License\n\nMIT\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftravisbreaks%2Fsamaritan","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftravisbreaks%2Fsamaritan","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftravisbreaks%2Fsamaritan/lists"}