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.
npx skills add AIDevGTM/gtm-cofounder --skill 19-research-the-marketgit clone --depth 1 https://github.com/AIDevGTM/gtm-cofounderWrote 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/aidevgtm/gtm-cofounder/19-research-the-market)<a href="https://agentmods.dev/skills/aidevgtm/gtm-cofounder/19-research-the-market"><img src="https://agentmods.dev/badge/skills/aidevgtm/gtm-cofounder/19-research-the-market/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/aidevgtm/gtm-cofounder/19-research-the-market"><img src="https://agentmods.dev/badge/skills/aidevgtm/gtm-cofounder/19-research-the-market.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 | $0.00122 | $0.02491 |
| Opus 5.5 | $0.00049 | $0.00996 |
| Sonnet 5.5 | $0.00024 | $0.00498 |
| Haiku 4.5 | $0.00012 | $0.00249 |
Grade A, and why
research-the-market 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 19d 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 — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research the market (what you didn't tell me)
The brief is what you know. A plan built only from the brief can only rearrange your own answers back to you. This is the step where the agent goes out and finds what you don't know yet: where your users actually are, how products like yours got their first users, and the openings you haven't seen.
Use this when: you've just written the brief and want a plan grounded in the real market, not just your own read. Or you're about to pick channels or plan a launch and "post on Reddit and HN" is the best idea on the table. Or the roadmap you got back felt like a tidy summary of what you said.
The core idea
The interview pulls out what's in your head. Research brings in what isn't: the specific subreddit where your ICP complained about this exact problem last week, the newsletter they all read, the three products in your space that launched in the last two years and what actually brought their first users, the competitor you didn't name.
Research is planning-grade evidence, not proof. Public evidence that 40 people asked for this in a GitHub issue is strong enough to plan a channel around. It is not strong enough to call the pain validated or to set a price. So it gets its own tag, [researched], between [assumption] and [validated]. Only a real conversation with a real user who is not a friend earns [validated].
Before you start: can your agent see the web?
- If your agent can browse or search the web, it runs the research for real, and every finding carries a link.
- If it can't, it says so up front and hands you a short list of exact searches to run (8 to 12, each a copy-paste query with what to look for). You paste back what you find, and it does the analysis. Same output, you're the browser.
Never fake it. An agent without web access that writes up "communities" from memory is guessing, and invented member counts and dead links destroy trust fast.
How to run this (for the agent)
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.
- 19d ago First seen · 133 lines · 122 tokens per session scan A 327bafe53904
research-the-market is a skill published in the GitHub repository AIDevGTM/gtm-cofounder (333 stars, last pushed 2d ago), licensed MIT. It adds 122 tokens to every session and 2,491 once invoked, about $0.0005 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 2026-10-17.
Other skills, from other repositories
git-phase-restore
Autonomous Git-based project phase restoration. Uses Git history (commits, tags, branches, diffs, semantic messages) to identify and restore any development phase automatically. Trigger for: "restore to when X worked", "go back to before Y broke", "show project phases", "undo the last feature", "roll back to [phase]"…
github-actions-trigger
Trigger GitHub Actions workflows from Claude Code. Auto-run tests, deploy, or validate when skills like run-pre-commit-checks complete.
e2b-sandboxes
E2B open-source cloud sandboxes for executing AI-generated code securely. USE FOR: run code in sandbox, e2b, code interpreter, execute AI code, secure code execution, isolated environment, AI coding agent execution, run untrusted code, cloud sandbox, code execution API, python sandbox cloud, AI agent code runner, safe…
screenshot-automation
Automatically captures, crops, and beautifies screenshots from URLs for your portfolio or docs.
exploratory-data-analysis
Performs bounded, local exploratory analysis of explicitly supported scientific files. Supports redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain…
webhook-management
Configure and validate CCAM webhook targets across supported chat, incident, automation, and generic providers. Use when listing provider requirements, creating or updating a target, scoping it to alert rules, sending a test notification, reviewing delivery history, or deleting a target.