Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/cofin/flowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/cofin/flow/flow)<a href="https://agentmods.dev/rules/cofin/flow/flow"><img src="https://agentmods.dev/badge/rules/cofin/flow/flow/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/cofin/flow/flow"><img src="https://agentmods.dev/badge/rules/cofin/flow/flow.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01177 | $0.01177 |
| Opus 5.5 | $0.00471 | $0.00471 |
| Sonnet 5.5 | $0.00235 | $0.00235 |
| Haiku 4.5 | $0.00118 | $0.00118 |
Grade A, and why
flow scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 19d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 64 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Flow Operational Rule
Activation: The always-applied repository rule and AGENTS.md activate Flow.
When a repository has .agents/, read its configured root, bundle index, active
spec, authoritative task worksheet, and applicable recursively nested knowledge
before acting. Operational skills resolve only from .agents/skills/.
Route through the flow skill and exactly one lifecycle skill. Refine plans
until every worksheet is executable, follow the selected worksheet without
improvising, and route contradictions through revise/refine. Rewrite revised
worksheets in place in present tense without historical transcripts, and keep
Markdown checklist markers ([x], [~], [ ]) synchronized with task state.
Apply direct Markdown frontmatter state synchronization and reconcile task-first.
Close and archive completed specs on the working branch without waiting for branch
merge; when scope grows beyond the active spec, finish the current spec and open a
follow-up spec via /flow:plan. Before deleting any spec during archive,
curate evidence-backed OKF chapters across nested domain, architecture, pattern,
testing, and gotcha folders (knowledge/index.md).
Write direct, zero-slop prose and code: avoid AI puffery, negative parallelisms,
formulaic transitions, and inline # comments in Python (use docstrings and PEP
585 types). Proactively fix adjacent lint, typing, or dead-code issues in touched
in-scope files and suggest the exact canonical /flow:<action> command next.
Use structured-choice-v1 for unresolved decisions: one decision at a time,
only through a currently allowed compatible native tool or the equivalent
sequential-text fallback. Run correctness review and then the mandatory fresh
quality review before finish/archive.
Commits remain local unless the user explicitly asks for delivery. Never create, move, force-update, or delete Git tags.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 19d ago Changed · +13 lines · +193 tokens per session 9479f6548adf
- 1mo ago First seen · 51 lines · 984 tokens per session scan A 1d8e082a8909
flow is a cursor rule published in the GitHub repository cofin/flow (19 stars, last pushed 20d ago), licensed Apache-2.0. It adds 1,177 tokens to every session, about $0.0047 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
mmr-review
Run multi-model code review with the MMR CLI (mmr review) before merging or finishing a change, or a multi-model design critique (mmr critique) of a design doc, plan, or proposed approach before building it. Use when the user asks to review code, a PR, a diff, or staged changes, or to critique/second-opinion a design…
cursor
Static reference for scaffold pipeline ordering, dependencies, and phase structure. Use ONLY for questions about pipeline design, step ordering, or dependency constraints — NOT for status, progress, or "what's next" queries (those go through scaffold-runner).
augx
Spec-driven development with openspec/changes/ (drafts) and openspec/archive/ (completed).
windsurfrules
The Beads MCP server (beads) provides real-time access to project tasks. Query tasks/list, tasks/get, dependencies/list, status/get for current state. No live task data is embedded in this file.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.