predict

A pre-implementation review command for high-stakes gflow-cli proposals. It asks five specialist perspectives to assess the proposal and returns a GO, CAUTION, or STOP decision with a confidence score.

In plain words
What is it for?
Use it before adding transports, changing authentication or selectors, migrating schemas, or starting a PLAN.md investigation task.
Why use it?
It exposes security, performance, architecture, usability, and other risks before code is written.

Command for Claude Code

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/ffroliva/gflow-cli/predict
Clone the repo
git clone --depth 1 https://github.com/ffroliva/gflow-cli

Made for: Claude Code.

Per session 62 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 307 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 $0.00062 $0.00307
Opus 5 $0.00031 $0.00153
Sonnet 5 $0.00012 $0.00061
Haiku 4.5 $0.00006 $0.00031

Measured yesterday against content hash 2d5421c30a16, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

predict 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 yesterday.

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.

.claude/commands/gflow/predict.md · 26 lines

What it actually says

/gflow:predict [proposal]

Read skills/predict/SKILL.md and follow its protocol now, passing $ARGUMENTS as the proposal description.

Do not call Skill(skill="predict") — the repo's skills/*/SKILL.md files are plain Markdown, not registered as Skill-tool-invocable (only .claude/commands/gflow/* are). Invoking it errors with Unknown skill: predict. Read the file directly instead.

The skill at skills/predict/SKILL.md runs five independent expert personas (Architect · Security/reCAPTCHA · Performance/Playwright · CLI UX · Devil's Advocate), resolves conflicts, and returns a GO / CAUTION / STOP verdict with a confidence score.

Typical workflow after a GO or CAUTION:

/gflow:scenario <feature>   →  edge cases + BDD skeleton
/gflow:plan <feature>       →  writes PLAN.md task checklist
/gflow:status               →  surfaces next task during execution
/gflow:check                →  before each commit
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. yesterday First seen · 26 lines · 62 tokens per session scan A 2d5421c30a16

Subscribe to this mod's changes

predict is a command published in the GitHub repository ffroliva/gflow-cli (136 stars, last pushed yesterday), licensed MIT. It adds 62 tokens to every session and 307 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.