{"id":51432628,"url":"https://github.com/alexsglife-re/multi-agent-working-group","last_synced_at":"2026-07-05T05:01:35.004Z","repository":{"id":369380996,"uuid":"1263539789","full_name":"alexsglife-re/multi-agent-working-group","owner":"alexsglife-re","description":null,"archived":false,"fork":false,"pushed_at":"2026-07-05T02:53:26.000Z","size":578,"stargazers_count":3,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"main","last_synced_at":"2026-07-05T03:11:06.051Z","etag":null,"topics":["agentic-workflows","ai-agents","ai-coding","claude-code","codex","developer-tools","hermes-agent","multi-agent","openclaw","openspec","videcoding","workflow"],"latest_commit_sha":null,"homepage":"","language":"Shell","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/alexsglife-re.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":"SECURITY.md","support":null,"governance":null,"roadmap":"docs/ROADMAP.md","authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-06-09T03:40:05.000Z","updated_at":"2026-07-05T02:53:30.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/alexsglife-re/multi-agent-working-group","commit_stats":null,"previous_names":["alexsglife-re/multi-agent-working-group"],"tags_count":4,"template":false,"template_full_name":null,"purl":"pkg:github/alexsglife-re/multi-agent-working-group","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alexsglife-re%2Fmulti-agent-working-group","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alexsglife-re%2Fmulti-agent-working-group/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alexsglife-re%2Fmulti-agent-working-group/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alexsglife-re%2Fmulti-agent-working-group/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/alexsglife-re","download_url":"https://codeload.github.com/alexsglife-re/multi-agent-working-group/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alexsglife-re%2Fmulti-agent-working-group/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35143802,"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":"2026-07-05T02:00:06.290Z","response_time":100,"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":["agentic-workflows","ai-agents","ai-coding","claude-code","codex","developer-tools","hermes-agent","multi-agent","openclaw","openspec","videcoding","workflow"],"created_at":"2026-07-05T05:01:34.489Z","updated_at":"2026-07-05T05:01:34.998Z","avatar_url":"https://github.com/alexsglife-re.png","language":"Shell","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003ch1 align=\"center\"\u003eMulti-Agent Working Group\u003c/h1\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"assets/social-preview.jpg\" alt=\"Multi-Agent Working Group - Safer multi-agent workflows for AI coding\" width=\"100%\"\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cstrong\u003eControlled multi-agent work for tasks where review, evidence, and safe exits matter more than speed.\u003c/strong\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"#quick-start\"\u003e\u003cimg src=\"https://img.shields.io/badge/Quick_Start-2_min-blue?style=for-the-badge\" alt=\"Quick Start\"\u003e\u003c/a\u003e\n  \u003ca href=\"docs/ADAPTERS.md\"\u003e\u003cimg src=\"https://img.shields.io/badge/Adapters-Codex_reference-black?style=for-the-badge\" alt=\"Codex reference adapter\"\u003e\u003c/a\u003e\n  \u003ca href=\"docs/VALIDATION.md\"\u003e\u003cimg src=\"https://img.shields.io/badge/Validation-local_checks-green?style=for-the-badge\" alt=\"Local validation\"\u003e\u003c/a\u003e\n  \u003ca href=\"LICENSE\"\u003e\u003cimg src=\"https://img.shields.io/badge/License-MIT-yellow?style=for-the-badge\" alt=\"MIT License\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\nAI agents are fast, but fast work can blur who reviewed what, which evidence was\nchecked, and whether a commit, push, release, or handoff was actually\nauthorized.\n\nMulti-Agent Working Group is a portable workflow protocol for AI-assisted work.\nIt gives careful tasks clear roles, independent review, durable evidence, and\nexplicit gates for validation, handoff, git, and release decisions.\n\nUse it when you want agents to help with serious work without silently approving\ntheir own output, losing context across conversations, or turning \"looks done\"\ninto \"safe to ship.\"\n\nThis repository ships Codex as the reference packaged adapter. The protocol is\ndesigned to be adapted to Claude, OpenClaw, Hermes Agent, and other agent\nruntimes, but non-Codex runtimes are adapter guidance or compatible patterns\nuntil their adapters are validated in real use.\n\n| Without a working protocol | With Multi-Agent Working Group |\n| --- | --- |\n| Reviews, handoffs, and git exits depend on memory and chat history. | Leader, PM, Advisor, Worker, and Reviewer roles produce explicit evidence. |\n| Agent output can look authoritative before it is checked. | Agent output is treated as evidence until the Leader verifies it. |\n| Commit, push, release, and cleanup boundaries can blur together. | Local completion, normal git gates, and Owner-only exclusions stay separate. |\n\nThe skill is intentionally conservative. It keeps the Leader responsible for\norchestration and verification, treats agent output as evidence rather than\nauthority, and separates local completion, normal git gates, and Owner-only\nexclusions.\n\n\u003e Current local version: `v0.4.15`. `v0.4.15` Fast Path And Anchor Guardrails is in progress on top of the public `v0.4.14` Adoption Scenarios And Adapter Guardrails release. Public release tags should point at reviewed commits; documentation version text alone is not a release, deployment, or external publication claim. For now, version tracking lives in `README.md`, `CHANGELOG.md`, and release tags when they are created, while `agents/openai.yaml` remains versionless interface metadata.\n\n## Quick Start\n\nUse the protocol when a task benefits from independent critique, delegated work,\ncontrolled commit or release gates, cross-conversation continuity, or a clear\nrecord of what was checked before moving forward.\n\nCodex reference-adapter minimal install:\n\n```sh\ngit clone https://github.com/alexsglife-re/multi-agent-working-group.git\nmkdir -p ~/.codex/skills/multi-agent-working-group\nmkdir -p ~/.codex/skills/multi-agent-working-group/references\ncp multi-agent-working-group/SKILL.md ~/.codex/skills/multi-agent-working-group/SKILL.md\ncp -R multi-agent-working-group/references/. ~/.codex/skills/multi-agent-working-group/references/\n```\n\nTypical prompt:\n\n```text\nUse the multi-agent-working-group skill for this task.\n```\n\nFor full checkout use, migration, validation, and scenario guidance, see\n`docs/INSTALLATION.md` and `docs/ADOPTION.md`.\n\n## Use Cases\n\nUse Multi-Agent Working Group when you want to:\n\n- coordinate PM, Advisor, Worker, and Reviewer roles during development;\n- review whether an open-source repository is ready for promotion;\n- improve the README first screen and quickstart path;\n- draft release notes, Show HN posts, Reddit variants, LinkedIn posts, or X\n  threads without publishing them automatically;\n- create a staged launch or release-preparation plan;\n- check whether public copy overstates runtime support, automation, safety, or\n  maturity;\n- run OpenSpec-backed work through C0 goal analysis, C1 proposal, C2\n  implementation, C3 closeout, and C4 archive;\n- preserve compact handoff evidence across long-running or spec-bound work;\n- keep normal commit and push gates separate from Owner-only release,\n  deployment, publication, force-push, destructive, secret, auth, and permission\n  actions.\n\nFor scenario-specific adoption guidance, see `docs/ADOPTION.md`.\n\n## Safety Boundaries\n\nThis protocol gives agents a workflow and evidence structure. It does not, by\nitself:\n\n- authorize commits, pushes, releases, tags, deployments, force-pushes,\n  destructive operations, or external publication;\n- edit GitHub metadata such as repository description, topics, homepage URL, or\n  social preview;\n- make Advisor, PM, Worker, Reviewer, handoff, memory, or template output\n  authoritative before Leader verification;\n- replace project tests, CI, secret scanning, code review, or OpenSpec archive;\n- read secrets, credentials, Keychain data, browser data, or unrelated projects;\n- automatically create successor conversations, spawn Workers, measure diff\n  size, close role agents, repair cleanup failures, or install external tools;\n- make every named runtime fully supported. Runtime adapters must be validated\n  before the project describes them as supported.\n\n## Repository Layout\n\n| Path | Purpose |\n| --- | --- |\n| `SKILL.md` | Codex reference-adapter entry point, hard-boundary summary, and progressive-reference router. |\n| `references/` | Progressive reference files that expand `SKILL.md` without weakening its constraints. |\n| `agents/openai.yaml` | Agent-facing metadata used by the Codex adapter bundle. |\n| `LICENSE` | MIT license for public reuse. |\n| `CONTRIBUTING.md` | Issue, pull request, and validation guidance for contributors. |\n| `SECURITY.md` | Sensitive-content and vulnerability-reporting guidance. |\n| `CODE_OF_CONDUCT.md` | Basic community participation expectations. |\n| `CHANGELOG.md` | Local version and stabilization notes. |\n| `docs/ROADMAP.md` | Development priorities and staged project direction. |\n| `docs/INSTALLATION.md` | Local installation, global skill sync, migration, and adoption guidance. |\n| `docs/ADAPTERS.md` | Platform adapter status, maturity labels, and runtime mapping checklist. |\n| `docs/ROLE_BOUNDARIES.md` | Working notes for Leader direct execution and role separation. |\n| `docs/VALIDATION.md` | Checklist for reviewing changes before publication. |\n| `examples/` | Copyable workflow examples, including blocked, git-gate, and OpenSpec C0-C4 scenarios. |\n| `templates/` | Fill-in templates for common role outputs, handoffs, blocked reports, and git gates. |\n| `openspec/` | OpenSpec changes, archived changes, and accepted specs. |\n| `scripts/validate-local.sh` | Lightweight read-only local validation command for repo docs, OpenSpec, and installed skill sync. |\n\nThe `openspec/changes/archive/` directory is design history. It is useful when\nyou want to understand why a rule exists, but ordinary users do not need to read\nthe archived changes before installing or using the skill.\n\n## License\n\nThis project is released under the MIT License. See `LICENSE`.\n\n## Adoption Docs\n\nStart with these docs when you need more than the quick path above:\n\n- `docs/ADOPTION.md`: scenario guide for documentation, release preparation,\n  long-running work, handoff, and ordinary small tasks.\n- `docs/INSTALLATION.md`: local checkout use, Codex reference-adapter install,\n  global skill sync, and migration boundaries.\n- `docs/ADAPTERS.md`: adapter maturity labels, mapping checklist, and future\n  runtime guide template.\n- `docs/VALIDATION.md`: validation checklist for docs, OpenSpec, release\n  readiness, and local checks.\n\nFor small read-only or low-risk documentation tasks, the Leader may complete the\nwork directly when every Small Task Mode condition in `SKILL.md` is met. Commit,\npush, archive, tag, release, deployment, publication, and other gated actions\nstill use the stricter rules.\n\n## Copyable Templates\n\nUse the templates in `templates/` when a workstream needs a consistent fill-in shape for C0 analysis, PM review, Advisor review, Worker assignment, Worker return, Reviewer report, blocked report, compact handoff, successor startup packet, or commit/push gate evidence.\n\nThe v0.4.13 templates are structure only. A filled template is evidence, not authorization. It does not replace PM, Advisor, Reviewer, validation, secret scanning, OpenSpec archive, CI/status, commit gates, push gates, release approval, cleanup-impact judgment, or Owner-only default-exclusion approval. Completion summaries and next-goal recommendations are reporting aids only; they do not authorize starting new work unless the Owner has already given explicit current-session authorization.\n\nWhen older v0.3 or earlier handoff documents have already grown large, preserve them as historical evidence. Create a new `templates/compact-handoff.md` current state card, copy only verified current facts into it, and reference the old document through the evidence index instead of appending or rewriting the old text.\n\n## Local Validation\n\nRun the local validation command before normal commit or push gates:\n\n```sh\n./scripts/validate-local.sh\n```\n\nAfter an OpenSpec-backed change is archived, use closeout mode:\n\n```sh\n./scripts/validate-local.sh --require-no-active-changes\n```\n\nThe command is read-only and does not use the network. It checks `SKILL.md` frontmatter, current version markers, accepted OpenSpec specs, `openspec validate --all`, active-change state, template and reference anchors, and the installed global skill plus required references when present. It is only a convenience check; it does not replace PM, Advisor, Reviewer, secret scanning, OpenSpec archive, git gate requirements, or runtime compliance with cleanup/delegation behavior.\n\n## Development Principles\n\n- Keep the skill readable before making it comprehensive.\n- Prefer explicit gates over implicit trust.\n- Make risk levels and stop conditions clear in plain language.\n- Add examples before adding more rules when examples would reduce ambiguity.\n- Treat continuity files, handoffs, and previous agent output as evidence, not authority.\n- Refresh long active handoffs around current verifiable state instead of appending old handoff text indefinitely.\n- Keep local validation lightweight and read-only until heavier automation is explicitly accepted.\n- Keep Leader direct execution narrow and visible; use `docs/ROLE_BOUNDARIES.md` before promoting new role rules into `SKILL.md`.\n\n## Validation\n\nBefore changing `SKILL.md`, review `docs/VALIDATION.md`. At minimum, confirm that:\n\n- The frontmatter remains valid.\n- Role boundaries remain clear.\n- Advisor permissions stay bounded.\n- Commit and push authorization rules are not weakened accidentally.\n- New instructions do not conflict with existing project or owner rules.\n- Public-facing docs do not include private machine paths, credentials, or\n  personal model-routing defaults.\n- Platform-facing docs do not claim full runtime support before an adapter is\n  validated.\n\nBefore a public release, also run:\n\n```sh\nopenspec validate --all\n./scripts/validate-local.sh --skip-global-skill\nrg -n \"(token|api[_-]?key|secret|password|private key|/Users/|gmail|Keychain|GITHUB_PAT)\" .\n```\n\nThe content scan is broad by design. Safety-instruction matches are expected;\nreal secrets, credentials, private project details, and local machine-specific\npaths must be removed before publication.\n\n## Current Status\n\nThis repository is in a documentation-first stabilization stage. Stage 1 foundation docs are mostly complete. The `v0.4.0` through `v0.4.14` stabilization and public-release preparation work is complete. `v0.4.15` clarifies Fast Path reading order and anchor guardrails while preserving `SKILL.md` as the fail-closed hard-boundary summary and mandatory router.\n\n`v0.4.15` is the current documented version. A future public release should be\ncreated only after the release-preparation diff is reviewed, validation passes,\nand an explicitly authorized tag points at the reviewed commit. Normal non-high-risk\ncommits and pushes follow the PM plus Advisor gate in `SKILL.md` with required\nevidence; high-risk and default-exclusion actions still require explicit Owner\napproval.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falexsglife-re%2Fmulti-agent-working-group","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Falexsglife-re%2Fmulti-agent-working-group","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falexsglife-re%2Fmulti-agent-working-group/lists"}