Skill

gcoc-research

Performs web research via GitHub Copilot CLI using free gpt-4.1 model. Use for simple topics, quota exhaustion, or explicit 'gcoc research' requests to avoid premium costs.

Bash

Node

developer-tools

npx claudepluginhub takazudo/claude-resources

Tool Access

This skill is limited to using the following tools:

Bash(bash *)Bash(timeout *)Bash(gtimeout *)Bash(node *)

Preview

Same behavior as [/gco-research](../gco-research/SKILL.md) — web research synthesized by Claude Code — but forces `gpt-4.1` from the start. No Premium Requests consumed.

SKILL.md

Similar Skills

cache-components

139.3k

Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.

cache-components

mcp-builder

124.2k

Guides building MCP servers enabling LLMs to interact with external services via tools. Covers best practices, TypeScript/Node (MCP SDK), Python (FastMCP).

9 files

anthropics-skills-13

Stats

Stars7

Forks0

Last CommitMay 7, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

GCOC Research (Cheap)

Same behavior as /gco-research — web research synthesized by Claude Code — but forces gpt-4.1 from the start. No Premium Requests consumed.

How It Differs from /gco-research

Model: gpt-4.1 (forced via GCO_MODEL=gpt-4.1)
Skip the GCO_USED_FALLBACK= stderr check — it's the user's deliberate choice, not a fallback

Process

Follow the exact same process as /gco-research.

Step 1: Understand the Research Topic

Clarify what needs to be researched — specific question, scope, how findings will be used.

Step 2: Build Research Prompt

Same research prompt as /gco-research (topic, specific questions, request for findings with sources/URLs, alternatives, recommendations, code examples).

Step 3: Run Copilot with forced cheap model

LOGDIR=$(node $HOME/.claude/scripts/get-logdir.js)
mkdir -p "$LOGDIR"
DATETIME=$(date +%Y%m%d_%H%M%S)
SLUG="<short-topic-slug>"

GCO_MODEL=gpt-4.1 \
  bash $HOME/.claude/skills/gco/scripts/gco-run.sh \
  "<research prompt>" \
  "$LOGDIR/${DATETIME}-gcoc-research-${SLUG}.md" \
  "$LOGDIR/${DATETIME}-gcoc-research-${SLUG}-stderr.log"

Timeout: 15 minutes.

Step 4: Collect Results

Read the output file
If empty or missing, check stderr log for errors
If Copilot timed out or failed, jump to Fallback
Do NOT check for GCO_USED_FALLBACK= marker — the cheap model is intentional

Step 5: Fallback

Same as /gco-research — if Copilot timed out, produced no output, or is not installed, spawn a researcher subagent.

Step 6: Synthesize

Organize findings, verify key claims where possible, present with source references.

Step 7: Save Research Log

node $HOME/.claude/scripts/save-file.js "{logdir}/{timestamp}-research-{topic}.md" "<content>"
pnpm dlx @takazudo/mdx-formatter --write <file>

Important Notes

Copilot does native web research — no need to paste content into the prompt
All file writing done by Claude Code, never by Copilot
gpt-4.1 has zero Premium multiplier on the Pro plan — safe to use when quota is tight
NEVER use ~ in paths — use $HOME