deepwork

A workflow for large or risky coding tasks that span several phases and files. It keeps planning, reviews, delegated work, and progress records together.

In plain words
What is it for?
Use it to plan and carry out broad code changes with review checkpoints and persistent progress tracking.
Why use it?
It reduces the chance of losing track of work or making unsafe changes during a long implementation or refactor.

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/mazumba/opencode-dockerized/deepwork
Any agent
npx skills add mazumba/opencode-dockerized --skill deepwork
Clone the repo
git clone --depth 1 https://github.com/mazumba/opencode-dockerized

Made for: Claude Code, Codex.

Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 964 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 $0.00038 $0.00964
Opus 5 $0.00019 $0.00482
Sonnet 5 $0.00008 $0.00193
Haiku 4.5 $0.00004 $0.00096

Measured yesterday against content hash 14c91ef8dca4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

deepwork 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.

.opencode/config/skills/deepwork/SKILL.md · 112 lines

How it starts

The opening of the file, as written. The whole thing — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Deepwork

Deepwork is an orchestrator workflow for heavy coding sessions. Use it when the work is broad, risky, multi-file, or likely to span several implementation phases. Do not use it for trivial edits, quick docs changes, or simple bug fixes.

Core Contract

When deepwork is active, the orchestrator must manage the work as a scheduler, not as the default implementation worker.

Required behavior:

  • keep OpenCode todos aligned with the active deepwork phase;
  • create and maintain a local markdown progress file under .slim/deepwork/;
  • write valuable research findings into that file as confirmed research context when they are received and reconciled;
  • draft a plan before implementation;
  • ask @oracle to review the plan and revise it until acceptable;
  • create a phased implementation/delegation plan;
  • before oracle reviews, add relevant confirmed research findings and file references to the deepwork file so oracle can review the plan or phase from accepted context instead of redoing discovery;
  • ask @oracle to review that implementation plan before execution;
  • after oracle review and before each implementation phase, decide the execution path: what can run in parallel, what must be sequential, which specialists to delegate to, and whether to split the same agent into multiple bounded lanes;
  • after each phase, validate, update the deepwork file, prepare the plan file for oracle review and ask @oracle to review the phase result, fix actionable issues, then continue;
  • when a phase includes @designer, preserve designer intent across later phases. Use @fixer only for mechanical follow-up that does not alter the UI/UX;
  • finish with final validation and a concise summary.

Designer Handoff Guardrail

When a deepwork phase includes @designer, treat the delivered UI/UX as accepted design intent for later phases. Record any important design decisions in the deepwork file before continuing.

After designer work:

  • preserve layout, rhythm, hierarchy, motion, spacing, color, affordances, responsiveness, and component feel;
  • review and improve user-facing copy with grounded, normal wording, but do not change visual structure or interaction intent;
  • route follow-up visual, responsive, motion, hierarchy, polish, or component-feel changes back to @designer;
  • use @fixer only for bounded mechanical follow-up that preserves the design exactly, such as wiring, tests, type fixes, or non-visual behavior changes;
  • if design intent must change, record why in the deepwork file before changing it.

Read the full file on GitHub · 112 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 · 112 lines · 38 tokens per session scan A 14c91ef8dca4

Subscribe to this mod's changes

deepwork is a skill published in the GitHub repository mazumba/opencode-dockerized (5 stars, last pushed 4d ago), licensed MIT. It adds 38 tokens to every session and 964 once invoked, about $0.0002 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.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

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…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

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…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens