haiku-refine

haiku-refine is a skill for Claude Code from gigsmart/haiku-method. It costs 18 tokens per session (211 once invoked), scanned A, original, Apache-2.0.

A process for changing an active task's requirements, a work unit, or an earlier stage's result while work is underway.

In plain words
What is it for?
Refining active intent, stage outputs, or units, recording feedback, and continuing the task lifecycle.
Why use it?
It lets corrections create follow-up work and rerun affected approvals without altering completed work.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

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

Good fit Refining active intent, stage outputs, or units, recording feedback, and continuing the task lifecycle.

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

README.md
[![agentmods](https://agentmods.dev/badge/skills/gigsmart/haiku-method/haiku-refine.svg)](https://agentmods.dev/skills/gigsmart/haiku-method/haiku-refine)
Your own site
<a href="https://agentmods.dev/skills/gigsmart/haiku-method/haiku-refine"><img src="https://agentmods.dev/badge/skills/gigsmart/haiku-method/haiku-refine.svg" alt="Measured on agentmods" height="20"></a>
Per session 18 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 211 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 $0.00018 $0.00211
Opus 5 $0.00009 $0.00105
Sonnet 5 $0.00004 $0.00042
Haiku 4.5 $0.00002 $0.00021

Measured 8d ago against content hash ae48766f1e50, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

haiku-refine 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 8d 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-refine/SKILL.md · 13 lines

What it actually says

Refine

  1. Call haiku_intent_list to find the active intent.
  2. Use ask_user_visual_question to ask what to refine: intent-level spec, specific unit, or upstream stage output.
  3. For stage-scoped refinement: use haiku_studio_stage_get and haiku_unit_list to read current state. Create new units via haiku_unit_write for the additional work — do NOT mutate completed units (forward-only invariant). The cursor picks up new units automatically (units with empty iterations[] are wave-ready).
  4. To target specific approvals for re-run, log a haiku_feedback finding with targets.invalidates: [<role>] against the affected unit. The terminal fix-hat clears those approvals on closure and the cursor reroutes through them.
  5. Call haiku_run_next to drive the lifecycle.
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. 8d ago First seen · 13 lines · 18 tokens per session scan A ae48766f1e50

Subscribe to this mod's changes

haiku-refine is a skill published in the GitHub repository gigsmart/haiku-method (24 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 18 tokens to every session and 211 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.

Related

Other skills, from other repositories

concentrate-forces

A decision-making method for situations where several tasks compete for limited time, attention, computing power, or budget.

HughYau/qiushi-skill · 187 tokens

mass-line

A method for collecting feedback and facts from multiple people or sources, turning them into a plan, and checking the plan with those affected.

HughYau/qiushi-skill · 195 tokens

workflows

A workflow guide for tasks that need several kinds of reasoning in sequence, such as research, finding the main conflict, choosing a small starting point, and planning sustained work.

HughYau/qiushi-skill · 157 tokens

protracted-strategy

A long-term planning method for difficult work that cannot be completed quickly, divided into stages with steady progress over time.

HughYau/qiushi-skill · 211 tokens

redteam-mindset

Red-team operator discipline — the mindset corrections that separate offensive testing from defensive WAPT. Built from authorized red-team work where conservative defaults caused multiple findings to be missed and one to be incorrectly retracted. Use at the START of any red-team engagement and again whenever feeling…

uphiago/recon-skills · 94 tokens

password-spray-methodology

End-to-end password spray playbook. User enumeration, lockout detection, password pattern generation, spray execution across all protocols, error code differentials, and engagement discipline. Unifies M365/Entra, Okta, Exchange, Kerberos, SharePoint, XMLRPC, OIDC, and AD SMB/WinRM spraying into one methodology.

uphiago/recon-skills · 76 tokens