Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add dhofheinz/open-plugins/plugin install refineryWrote 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.
[](https://agentmods.dev/skills/dhofheinz/open-plugins/refine)<a href="https://agentmods.dev/skills/dhofheinz/open-plugins/refine"><img src="https://agentmods.dev/badge/skills/dhofheinz/open-plugins/refine/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.
<a href="https://agentmods.dev/skills/dhofheinz/open-plugins/refine"><img src="https://agentmods.dev/badge/skills/dhofheinz/open-plugins/refine.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00093 | $0.03248 |
| Opus 5.5 | $0.00037 | $0.01299 |
| Sonnet 5.5 | $0.00019 | $0.00650 |
| Haiku 4.5 | $0.00009 | $0.00325 |
Grade A, and why
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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Refinery — Unified Specification Refinement
You are the orchestrator for the Refinery plugin. Your job is to interpret the user's intent and dispatch to the correct mode procedure, never to perform operations directly. Mode files (loaded on demand) contain the actual procedures.
Directive
Refinery treats specifications as typed nodes in an artifact graph. Every operation reads working-directory state from artifact frontmatter, executes a focused mode procedure, writes results back to artifacts, and exits. The plugin is stateless across invocations — all state lives in markdown files.
Your responsibility on every invocation:
- Parse
$ARGUMENTSto determine intent - Read working-directory state (read frontmatter only — do not load full artifact bodies until a mode requires them)
- Resolve to one of eleven modes
- Load the corresponding mode file via
Readfrom${CLAUDE_SKILL_DIR}/mode-<name>.md - Execute the mode's procedure (which may load further stage files, reference files, templates, and spawn agents)
- Report outcome and suggest next action
- Exit
You never embed mode logic directly in this orchestrator. You dispatch.
Working Directory
Default resolution order (highest priority first):
--output-dir=<path>flag on the invocation${user_config.working_directory}if set in plugin user configdocs/refinery/(built-in default, distinct fromdocs/specs/to avoid colliding with legacy artifacts)
If the resolved working directory does not exist and the user did not invoke init, prompt before creating it.
Input Parsing
Process $ARGUMENTS in the following order. First match wins.
- Empty arguments + working directory has artifacts → mode =
status - Empty arguments + working directory has no artifacts → AskUserQuestion: "What would you like to refine?" with options:
- "Initialize project conventions" →
init - "Start a new system from an idea" → prompt for idea then
advance --stage=principles - "Document an existing feature" → prompt for feature name then
advance --stage=feature-spec - "Cancel"
- "Initialize project conventions" →
- Mode keyword as first argument (
init,advance,iterate,review,finalize,check,tickets,update,status,archive,mark-implemented) → use that mode; remaining arguments are passed through to the mode file. --stage=<name>flag present → mode =advance; target stage =<name>- First argument is an existing file path (matches
*.mdand exists) → inspect file'sartifact:andstatus:frontmatter; suggest the most likely operation and confirm via AskUserQuestion if ambiguous. - First argument is a free-text idea (multi-word, no special characters that suggest a path or feature name) → mode =
advance; target stage =principles; idea text passed as input. - First argument is a single-word identifier that doesn't match a mode keyword and doesn't exist as a file → assumed to be a feature name; mode =
advance; target stage =feature-spec; scope =feature; feature name = the argument. - Otherwise → AskUserQuestion to clarify intent (do not guess).
What ships with it
34 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- mode-advance.md 7.8 KB
- mode-archive.md 6.5 KB
- mode-check.md 8.7 KB
- mode-finalize.md 7.4 KB
- mode-init.md 8.5 KB
- mode-iterate.md 12 KB
- mode-mark-implemented.md 9.6 KB
- mode-review.md 6.7 KB
- mode-status.md 5.5 KB
- mode-tickets.md 9.5 KB
- mode-update.md 6.3 KB
- references/agent-handoffs.md 12 KB
- references/commit-protocol.md 12 KB
- references/convergence.md 9.2 KB
- references/document-format.md 13 KB
- references/operation-bookkeeping.md 10.0 KB
- references/requirement-syntax.md 7.9 KB
- references/state-detection.md 7.1 KB
- references/ticket-format.md 12 KB
- stage-design.md 4.3 KB
- stage-feature-spec.md 6.7 KB
- stage-plan.md 4.8 KB
- stage-principles.md 5.8 KB
- stage-spec.md 4.2 KB
- stage-stack.md 5.6 KB
- templates/_conventions.md 4.5 KB
- templates/_glossary.md 1.3 KB
- templates/design.md 5.4 KB
- templates/feature-spec.md 4.1 KB
- templates/plan.md 4.2 KB
- templates/principles.md 5.0 KB
- templates/spec.md 4.8 KB
- templates/stack.md 4.6 KB
- templates/tickets.md 3.5 KB
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.
- 1mo ago First seen · 216 lines · 93 tokens per session scan A adc874559342
refine is a skill published in the GitHub repository dhofheinz/open-plugins (9 stars, last pushed 6mo ago), licensed MIT. It adds 93 tokens to every session and 3,248 once invoked, about $0.0004 per session on Opus 5.5. 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.
Other skills, from other repositories
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…