agentleFS
cofincofin/flow/AGENTS.md

flow

AGENTS.md19 starsChanged 5 days agoApache-2.0View on GitHub

Flow is a multi-harness Spec-Driven Development plugin for Google Antigravity, Claude Code, Codex CLI, Cursor, OpenCode V2, and VS Code / Copilot. Flow has no flow shell executable; invoke the flow skill or a /flow:* command. Initialize with the setup guide, then use .agents/bundles/index.md, .agents/bundles/knowledge/workflow.md, and .agents/skills/index.md. Execute one worksheet per subagent turn (model: inherit) and strictly one chapter per /flow:implement run (syncing present-tense OKF/docs and passing zero-context chapter review before stopping), preserve unrelated work, stage exact paths, and commit…

The file

Files
AGENTS.md

Flow Project Context

Flow is a multi-harness Spec-Driven Development plugin for Google Antigravity, Claude Code, Codex CLI, Cursor, OpenCode V2, and VS Code / Copilot. Flow has no flow shell executable; invoke the flow skill or a /flow:* command.

Architecture & Routing

  • Lifecycle Router: skills/flow/SKILL.md routes across flow-setup (setup), flow-planning (prd, spec [alias plan], revise [alias refine], research [alias task]), flow-execution (implement, status), flow-learn (learn [aliases resynthesize, refresh, sync]), flow-docs (docs [subcommands write, validate, capture, sync, pr]), and flow-completion (review, finish, archive, cleanup, revert, validate).
  • Foundation & Harness Skills: okf (OKF v0.2 knowledge bundles and provenance), debloat (deep-module deletion test, public-seam test strength, and anti-slop prose/context pruning), and multi-harness-plugins (project-local .agents/ and multi-harness skill, subagent, hook, rule, and MCP standards).
  • Stateful Planning vs. Subagent-Managed Chapters & Zero-Context Verification: Stateful Co-Designers/Adjudicators (@prd-orchestrator, @plan-generator) decompose multi-module work into context-bounded chapters (specs/<flow_id>/chapters/ch01-<slug>/..chNN-<slug>/, 1–4 vertical-slice tasks per chapter, zero upper cap on chapter count) authored upfront in parallel subagents without truncating engineering or documentation detail; Zero-Context Reviewers & Isolated Executors (@code-reviewer, @quality-reviewer, @researcher, @executor) operate under Zero-Context Isolation Rules, Bounded Read Manifests, Public-Seam Vertical-Slice TDD, and cross-harness Stop review/sync hooks.
  • Present-Tense Direct Markdown State Contract: Root specifications (.agents/bundles/specs/<flow_id>/spec.md), chapter catalogs (chapters/index.md), chapter specs (chapters/<ch_id>/spec.md), and task worksheets (tasks/<id>-<slug>.md) are kept synchronized in place in present tense—showing current verified state, remaining chapters, and the full plan without historical sediment (skills/flow/references/state.md).

Discipline

Initialize with the setup guide, then use .agents/bundles/index.md, .agents/bundles/knowledge/workflow.md, and .agents/skills/index.md. Execute one worksheet per subagent turn (model: inherit) and strictly one chapter per /flow:implement run (syncing present-tense OKF/docs and passing zero-context chapter review before stopping), preserve unrelated work, stage exact paths, and commit locally only after verification passes. Never create or mutate Git tags. Pushes and remote writes require explicit user instruction.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.