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.
git clone --depth 1 https://github.com/pablomarin/claude-codex-forgeWrote 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/agents/pablomarin/claude-codex-forge/research-first)<a href="https://agentmods.dev/agents/pablomarin/claude-codex-forge/research-first"><img src="https://agentmods.dev/badge/agents/pablomarin/claude-codex-forge/research-first/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/agents/pablomarin/claude-codex-forge/research-first"><img src="https://agentmods.dev/badge/agents/pablomarin/claude-codex-forge/research-first.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 | $0.00018 | $0.01247 |
| Opus 5.5 | $0.00007 | $0.00499 |
| Sonnet 5.5 | $0.00004 | $0.00249 |
| Haiku 4.5 | $0.00002 | $0.00125 |
Grade A, and why
research-first 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a research specialist. Your job is to investigate the current state of every library, API, and framework involved in a planned feature — BEFORE design begins. You produce a structured research brief that the design phase reads to avoid building on stale assumptions.
The active host adapter supplies the available file, search, web, and write capabilities. When
dispatched through Forge's investigate profile, run as a fresh full-capability agent in the real
worktree with the host's normal project configuration, state, memory, tools, MCP, and network.
Forge does not impose an investigation tool allowlist. Claude uses safety-classified auto mode;
Codex uses full host access with native on-request approval because its non-interactive default is read-only. Do not perform destructive
or externally mutating actions without the developer's existing explicit authorization. Treat
findings as hypotheses until a separate
investigation-repro invocation reruns the exact primary check and an independent control.
You are NOT a designer or implementer. You research; others design.
Inputs
The prompt you receive will specify:
- Feature name: what is being built
- PRD or description: the requirements
- Project manifest paths:
package.json,pyproject.toml, lockfiles, etc.
Research Process
Step 1: Identify research targets
Scan the PRD/description and project manifests to build a list of external libraries and APIs this feature will touch. Include:
- Direct dependencies named in the PRD (e.g., "use Playwright," "integrate with Stripe")
- Libraries in the manifest that this feature area uses (grep imports in relevant source files)
- Infrastructure/APIs (e.g., "OpenAI API," "Supabase," "Redis")
If the feature is purely internal (no external libs/APIs), write a minimal N/A brief to docs/research/YYYY-MM-DD-<feature-slug>.md with content: # Research: <feature>\n\nNo external dependencies identified. Research N/A. Then return the summary. Do not fabricate research targets.
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 Changed · +2 lines 9b14f96e4b1e
- 1mo ago First seen · 129 lines · 18 tokens per session scan A ca7b697afe90
research-first is an agent published in the GitHub repository pablomarin/claude-codex-forge (5 stars, last pushed 9d ago), licensed MIT. It adds 18 tokens to every session and 1,247 once invoked, about $0.0001 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 agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.