shipkit-direction

A project-direction workflow that defines why a project exists, who it serves, what it should provide, and how it will be built.

In plain words
What is it for?
Use it to start a project, define its direction, or establish its vision, users, product, engineering approach, and goals.
Why use it?
It turns a vague project idea or repository into a shared foundation while asking only for decisions that cannot be grounded in available information.

Skill for Claude CodeCodex

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 skills/stefan-stepzero/shipkit/shipkit-direction
Any agent
npx skills add stefan-stepzero/shipkit --skill shipkit-direction
Clone the repo
git clone --depth 1 https://github.com/stefan-stepzero/shipkit

Made for: Claude Code, Codex.

Per session 80 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,336 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 $0.00080 $0.01336
Opus 5 $0.00040 $0.00668
Sonnet 5 $0.00016 $0.00267
Haiku 4.5 $0.00008 $0.00134

Measured yesterday against content hash 52f5a0c366fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

shipkit-direction 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.

install/skills/shipkit-direction/SKILL.md · 110 lines

How it starts

The opening of the file, as written. The whole thing — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.

shipkit-direction — Define the Foundation

Purpose: The consumer-facing DIRECTION entry point. Takes a project (a repo, a one-line goal, or nothing) and drives the definitional foundation — vision → stage → users → product → engineering → goals — to a coherent bar, autonomously: it proposes everything it can ground in a cited signal and asks only the few high-leverage decisions.

shipkit-direction is a thin caller. It does not implement its own loop — it states the what (steps + bar + which steps are elicitive) and hands off to the engine (shipkit-orchestrate), which owns the how (dispatch → reconcile → re-dispatch → stop-at-bar, with elicitation bubbled to the main session). This replaces the previous forked direction loop.

Read first: install/shared/references/ground-or-ask-calibration.md (the propose-vs-ask rule), install/shared/references/elicitation-protocol.md (the marker/resume mechanics the elicitive steps use), and install/shared/references/core-automation.md (the engine contract).


What it does

Invoke shipkit-orchestrate in autonomous mode (propose variant) over the Direction steps. The engine runs inline in the main session, dispatches each step, reconciles its artifact against the bar, bubbles any elicitation marker to the user, and stops when the foundation is coherent.

Pass the engine the four inputs (calling interface in shipkit-orchestrate/SKILL.md):

Input For Direction
steps the definitional chain, in order: why-project → stage → product-discovery → product-definition → engineering-definition → product-goals → engineering-goals, with review-direction as the assess gate
bar each step's named .shipkit/ artifact exists AND is grounded (every field cited or flagged per the calibration), AND review-direction passes (coherent, no gaps)
mode autonomous — propose the grounded foundation, ask only the high-leverage few (fall back to interactive when signals are thin)
elicitive the steps that may pause for input: why-project, stage, product-discovery, product-goals, engineering-goals (plus feedback-bug on the feedback re-entry path — six in all). The product-definition / engineering-definition synthesis steps are not elicitive — they derive from upstream artifacts and only propose. Each elicitive step emits NEEDS_ELICITATION:<slug>; the engine batches the questions to the user and resumes.

Read the full file on GitHub · 110 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. yesterday First seen · 110 lines · 80 tokens per session scan A 52f5a0c366fe

Subscribe to this mod's changes

shipkit-direction is a skill published in the GitHub repository stefan-stepzero/shipkit (1 stars, last pushed 1mo ago), licensed MIT. It adds 80 tokens to every session and 1,336 once invoked, about $0.0004 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 2026-08-31.