flow

A command that turns a sprint's acceptance criteria into a human walkthrough. Acceptance criteria are statements describing what must be true for the work to count as complete.

In plain words
What is it for?
Use it to create manual verification steps for login flows, pages, routes, and other acceptance criteria that need human observation.
Why use it?
Automated checks cannot observe everything, especially actions a person must perform in a browser or terminal. The walkthrough explains what to open, click, or verify.

Command

Install

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.

agentmods
npx agentmods add commands/koolamusic/claudefiles/flow
Clone the repo
git clone --depth 1 https://github.com/koolamusic/claudefiles
Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 921 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00052 $0.00921
Opus 5 $0.00026 $0.00461
Sonnet 5 $0.00010 $0.00184
Haiku 4.5 $0.00005 $0.00092

Measured 2d ago against content hash 6a0f351fe5c8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

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 2d 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.

plugins/jira/commands/flow.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Translate the active sprint's acceptance predicates into steps a human can walk. Not a test harness — a walkthrough. The reader is a person with a browser, a terminal, and no context.

Position in the loop

Three sensors cover one goal set (features/*.feature): nyquist senses machine-testable predicates during execute; flow is the human sensor — it tells a person how to observe the rest; warden senses deterministically post-sprint. Flow's output also feeds /warden:design (Prerequisites, Identity scope).

Steps

  1. Locate the active sprint: slug=$(cat .jira/CURRENT). If empty or missing, stop.

  2. Read the sources:

    • .jira/sprints/<slug>/features/*.feature — the predicates (required; if missing, stop and point at /jira:plan)
    • All *-PLAN.md — implementation context per predicate (via effects:); --plan N narrows to one plan's predicates
    • CONTEXT.md — locked decisions that shape expected behavior
    • EXECUTION.md ## Nyquist results block, if present — which predicates are already machine-sensed
  3. Classify each predicate:

    • machine-sensed — a nyquist results line carries its ID with a green test
    • human-required — marked not_testable_in_code / covered-by-inspection, or no nyquist line at all
    • If EXECUTION.md doesn't exist yet (pre-execute), every predicate is human-required
  4. Write .jira/sprints/<slug>/FLOW.md:

    ---
    sprint: <slug>
    generated: <iso date>
    predicates: <total> (<human-required count> need human sensing)
    ---
    
    # Flow: <sprint goal>
    
    ## Before you start
    
    Services to run, accounts to use, data that must exist — concrete: the command
    that starts the server, the exact login email/role (never a password — reference
    where it's shared), the seed state.
    
    ## Verify by hand (predicates machines could not sense)
    
    ### TOK-02 — <scenario title>
    1. Log in as <role> at <route>
    2. <do the Given — set up the precondition>
    3. <do the When — the action, exact button/route/command>
    4. **Expect:** <the Then, observably — what you see when it holds, and what
       failure would look like>
    
    ## Already machine-sensed (spot-check optional)
    
    ### TOK-01 — <scenario title> — sensed by `tests/auth.test.ts`
    One-line human spot-check, for skeptics.
    

Read the full file on GitHub · 71 lines

Changes

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.

  1. 2d ago First seen · 71 lines · 52 tokens per session scan A 6a0f351fe5c8

Subscribe to this mod's changes

flow is a command published in the GitHub repository koolamusic/claudefiles (132 stars, last pushed 3d ago), licensed MIT. It adds 52 tokens to every session and 921 once invoked, about $0.0003 per session on Opus 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.