sunlight-deepresearch

sunlight-deepresearch is a skill for Claude Code, Codex from sunlight-research-ai/sunlight-skills. It costs 72 tokens per session (1,966 once invoked), scanned A, original, MIT.

A structured method for researching broad or uncertain questions with several agents working on separate parts. It combines their findings, checks evidence quality, and resolves disagreements before producing a report.

In plain words
What is it for?
Use it to investigate technical, academic, market, product, or strategic questions; compare options; verify claims; and map risks or unknowns.
Why use it?
It reduces the risk of relying on one incomplete search or accepting unsupported claims.

Skill for Claude CodeCodex

Written for Claude Code and Codex: shipped in a Claude Code plugin, but also agents/openai.yaml present. Also seen: mentions subagents; mentions Claude Code; mentions Codex.

Part of the sunlight-skills plugin — 1 skill shipped together

Good fit Use it to investigate technical, academic, market, product, or strategic questions; compare options; verify claims; and map risks or unknowns.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch
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 sunlight-research-ai/sunlight-skills --skill sunlight-deepresearch
Clone the repo
git clone --depth 1 https://github.com/sunlight-research-ai/sunlight-skills

Made for: Claude Code, Codex.

Or install sunlight-skills, the plugin that ships this one along with the rest of its 1 skill.

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 sunlight-deepresearch

README.md
[![agentmods](https://agentmods.dev/badge/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch/github.svg)](https://agentmods.dev/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch)
Your own site
<a href="https://agentmods.dev/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch"><img src="https://agentmods.dev/badge/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch/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 sunlight-deepresearch

Your own site · 80×15
<a href="https://agentmods.dev/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch"><img src="https://agentmods.dev/badge/skills/sunlight-research-ai/sunlight-skills/sunlight-deepresearch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 72 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,966 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 $0.00072 $0.01966
Opus 5 $0.00036 $0.00983
Sonnet 5 $0.00014 $0.00393
Haiku 4.5 $0.00007 $0.00197

Measured 12d ago against content hash 7532b800c63d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

sunlight-deepresearch 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 12d ago.

The scan reads SKILL.md. This mod also ships 3 executable files (scripts/create-research-plan.py, scripts/evaluate-source-coverage.py, scripts/search-providers.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

skills/sunlight-deepresearch/SKILL.md · 116 lines

How it starts

The opening of the file, as written. The whole thing — 116 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Sunlight Deep Research

Overview

Run deep research as a disciplined subagent orchestration workflow: decompose the question, send focused investigators, collect evidence, resolve conflicts, synthesize, and verify before final delivery.

When to Use

  • The user asks for deep research on a broad, ambiguous, technical, market, product, academic, or strategic topic.
  • The user wants a reusable research process rather than a one-pass answer.
  • The user asks to compare options, investigate claims, map a domain, produce a sourced report, or find risks and unknowns.
  • The user wants to run the workflow in Codex, Claude Code, OpenCode, or another agentic assistant.
  • The user wants to implement the workflow in an application; LangGraph may be used, but is not required.

When NOT to Use

  • Do not use for simple factual questions that can be answered directly with one reliable source.
  • Do not use when the user explicitly wants a direct answer without a research process.
  • Do not require LangGraph, a codebase, a database, or a hosted backend unless the user asks to build a durable app.

Contract

  • Do not answer broad research questions from a single pass when subagents are available.
  • Treat every sunlight-deepresearch run as thorough research; do not offer lighter modes.
  • Create file-backed research artifacts as the run progresses: brief, plan, source registry, per-source evidence files, subagent notes, evaluator audits, and final report.
  • Break the objective into independent research tracks before dispatching work.
  • Give each subagent a narrow brief, explicit output shape, and source/evidence requirements.
  • Require each investigator to complete a full query sweep and coverage check before stopping.
  • Register every useful fetched link in a source registry and create a per-source evidence file before using it in findings.
  • Ask subagents to separate evidence, inference, uncertainty, and open questions.
  • Synthesize only after reviewing the returned findings.
  • Run citation, source-quality, coverage, and contradiction evaluators against the actual final report before final delivery.
  • Resolve conflicts with targeted follow-up research instead of smoothing them over.
  • Before dispatching investigators, run python3 skills/sunlight-deepresearch/scripts/search-providers.py --check from the user's project or run folder to detect configured Linkup, Exa, and Tavily keys. If the skill path differs, use the installed skill's script path.
  • Run a final verifier or critic pass before presenting the final report.
  • Block final delivery when key findings or factual sentences lack linked sources.
  • State limitations, confidence, and unresolved questions in the final output.

Read the full file on GitHub · 116 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. 12d ago First seen · 116 lines · 72 tokens per session scan A 7532b800c63d

Subscribe to this mod's changes

sunlight-deepresearch is a skill published in the GitHub repository sunlight-research-ai/sunlight-skills (5 stars, last pushed 2mo ago), licensed MIT. It adds 72 tokens to every session and 1,966 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.

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

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…

microsoft/ai-agents-for-beginners · 200 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

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.

microsoft/vscode · 53 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

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…

vercel/next.js · 83 tokens