haiku-autopilot

haiku-autopilot is a skill for Claude Code from gigsmart/haiku-method. It costs 23 tokens per session (938 once invoked), scanned A, original, Apache-2.0.

An automated workflow that takes a software task from planning through building, review, and delivery. It reduces the number of points where a person must approve the next stage.

In plain words
What is it for?
Use it to run an H·AI·K·U intent through its full lifecycle from a description to delivery. It can also switch an existing intent into this mode.
Why use it?
It is for work where the usual human review gates can be handled automatically, while still stopping when outside approval or a final completion review is needed.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: mentions subagents.

Part of the haiku plugin — 31 skills, 2 hooks, 2 MCP servers shipped together

Good fit Use it to run an H·AI·K·U intent through its full lifecycle from a description to delivery. It can also switch an existing intent into this mode.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/gigsmart/haiku-method/haiku-autopilot
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.

Any agent
npx skills add gigsmart/haiku-method --skill haiku-autopilot
Clone the repo
git clone --depth 1 https://github.com/gigsmart/haiku-method

Made for: Claude Code.

Or install haiku, the plugin that ships this one along with the rest of its 31 skills, 2 hooks, 2 MCP servers.

Wrote 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.

agentmods badge for haiku-autopilot

README.md
[![agentmods](https://agentmods.dev/badge/skills/gigsmart/haiku-method/haiku-autopilot/github.svg)](https://agentmods.dev/skills/gigsmart/haiku-method/haiku-autopilot)
Your own site
<a href="https://agentmods.dev/skills/gigsmart/haiku-method/haiku-autopilot"><img src="https://agentmods.dev/badge/skills/gigsmart/haiku-method/haiku-autopilot/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.

agentmods 80×15 button for haiku-autopilot

Your own site · 80×15
<a href="https://agentmods.dev/skills/gigsmart/haiku-method/haiku-autopilot"><img src="https://agentmods.dev/badge/skills/gigsmart/haiku-method/haiku-autopilot.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 23 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 938 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00023 $0.00938
Opus 5 $0.00012 $0.00469
Sonnet 5 $0.00005 $0.00188
Haiku 4.5 $0.00002 $0.00094

Measured 9d ago against content hash 501e2b989eae, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

haiku-autopilot 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 9d 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.

plugin/skills/haiku-autopilot/SKILL.md · 36 lines

How it starts

The opening of the file, as written. The whole thing — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Autopilot

Run the full H·AI·K·U lifecycle autonomously from description to delivery. Autopilot is one of the three intent modes (discrete | continuous | autopilot). It tells the workflow engine to promote ask review gates to auto, so the lifecycle advances without human intervention on stage gates. External gates and the intent-completion review still pause (structural signals the workflow engine can't synthesize).

Process

  1. If no active intent exists, create one with /haiku:haiku-start, then set autopilot with haiku_select_mode { intent: "<slug>", modes: ["autopilot"] }. Passing a single mode auto-selects it (no picker pause) and writes mode: autopilot plus the studio's full stage list. Do NOT pass mode to haiku_intent_create — studio/mode/stage are engine-managed and start unset.
  2. For an existing intent, set the mode with haiku_select_mode { intent: "<slug>", modes: ["autopilot"] } — a single-option list auto-selects with no picker pause and writes mode (+ stages) through the one engine-sanctioned path. Do NOT use haiku_intent_set { field: "mode" } or haiku_human_writemode is an engine-managed field that both reject (intent_field_engine_only); haiku_select_mode is the ONLY mode setter. Do NOT set a separate autopilot: true boolean — that is a deprecated pattern. (If the intent has already started a stage, select_mode refuses transitions into/out of quick; surface that to the user.)
  3. Optional: skip the final intent review by calling haiku_intent_set { intent: "<slug>", field: "intent_completion_review", value: false }. Do NOT set this unless the user explicitly wants truly headless completion; the completion review is the bookend that prevents silent intent-completion on stage-gate pass.
  4. Drive the loop by calling haiku_run_next { intent: "<slug>" }. Repeat on every return. When a subagent returns, re-call haiku_run_next to advance.

What still pauses autopilot

  • External gates (external or compound [external, ask]). They need a real PR/MR merge signal and cannot be auto-approved.
  • await gates. Waiting for a non-review external event (customer response, pipeline, etc.).
  • Elicitation-required decisions. Design-direction picks, visual approvals.
  • Scope explosions (see guardrails below).
  • Intent-completion review (unless intent_completion_review: false is set on the intent).

Read the full file on GitHub · 36 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. 9d ago First seen · 36 lines · 23 tokens per session scan A 501e2b989eae

Subscribe to this mod's changes

haiku-autopilot is a skill published in the GitHub repository gigsmart/haiku-method (24 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 23 tokens to every session and 938 once invoked, about $0.0001 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.