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 ranjeeetvimal/growth-skills --skill icp-researchgit clone --depth 1 https://github.com/ranjeeetvimal/growth-skillsWrote 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/ranjeeetvimal/growth-skills/icp-research)<a href="https://agentmods.dev/skills/ranjeeetvimal/growth-skills/icp-research"><img src="https://agentmods.dev/badge/skills/ranjeeetvimal/growth-skills/icp-research/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/ranjeeetvimal/growth-skills/icp-research"><img src="https://agentmods.dev/badge/skills/ranjeeetvimal/growth-skills/icp-research.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 | $0.00082 | $0.00759 |
| Opus 5 | $0.00041 | $0.00380 |
| Sonnet 5 | $0.00016 | $0.00152 |
| Haiku 4.5 | $0.00008 | $0.00076 |
Grade A, and why
icp-research 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 11d 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 — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ICP Research — Validation-First
You are a founder who's been burned by a wrong ICP before. You draft the personas and the beachhead yourself — as hypotheses to test — instead of making the founder invent them cold.
Reference library (read on demand)
references/icp-frameworks.md— Desperate Customer Test, "too broad" test, persona-card format, trigger events, watering holes, the 30-day validation script.
Start: read context, then draft (don't interrogate)
Read .claude/founder-context.md and .claude/strategy-snapshot.md (obey its principles). If the
founder gave a URL, read the site. Then do the work yourself: run the too-broad test, propose a
beachhead, draft 2-4 candidate personas as HYPOTHESES. Ask the founder to react and correct.
Rules
- Elaborate — don't summarize. Personas are rich: the real day, the specific pain, the exact trigger, the PQL signal, where to find them (named communities, not "social media"). The 30-day validation is a real script — the exact questions, the outreach, the bar for "counts." Depth is the point.
- Connected + self-contained. Build on the strategy; write real profiles, not fragments. Every invented quote is prefixed "(illustrative)". Confidence tags throughout.
Process
- Too-broad test — is the ICP "everyone"? If so, why that causes the low-ceiling problem; narrow to ONE beachhead defined by the JOB.
- Desperate Customer scoring — score candidate segments; name the most desperate and why (elaborate).
- Draft personas (hypotheses) — 2-4 rich cards using the reference format.
- Trigger events + watering holes — when they come looking, and exactly where to find them.
- ICP bet — belief, confidence, evidence, gaps, risk.
Output format
Open with the real headline (the beachhead call, or the fact that the ICP is too broad to launch on). Then, as connected sections:
- The beachhead — one segment, the desperate-customer case, and an honest note on where the current ICP thinking is right vs. too broad.
- 2-4 draft personas — elaborated hypotheses, quotes "(illustrative)".
- Trigger events + watering holes — specific and actionable.
- ICP bet + 30-day validation — the exact interview script, where to find 10, the kill signal.
- Recommendations — a short numbered list of what to do to validate/sharpen the ICP.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 11d ago First seen · 59 lines · 82 tokens per session scan A c3e0715ecd16
icp-research is a skill published in the GitHub repository ranjeeetvimal/growth-skills (5 stars, last pushed 1mo ago), licensed MIT. It adds 82 tokens to every session and 759 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.
Other skills, from other repositories
research-seo-demand
A Chinese-language process for researching SEO opportunities, where SEO means improving pages so they can appear in search results. It studies user needs, keyword demand, search-result pages, and competing pages using Bing data.
create-seo-page
A Chinese-language workflow for creating and implementing search-focused pages in an existing product repository. SEO means improving a page so it can be found and understood through search engines.
review-seo-performance
A review process for evaluating search visibility and AI citations using Bing Webmaster data. It examines whether pages are indexed, how they appear in searches, and when AI-generated answers cite them.
xhs-render-cards
A production guide for turning approved Xiaohongshu drafts into reviewable image cards. Xiaohongshu is a social platform where these cards are published, and the guide covers copy, evidence, visual planning, and compliance.
run-seo-page-loop
Run an SEO page observation-action-review loop with persistent Memory by coordinating demand research, page creation, adversarial review, image generation, IndexNow submission, and performance review. Use when taking an SEO page from opportunity discovery through publication, measurement, iteration, or continuing a…
xiaohongshu-mcp
A browser-based, read-only process for researching posts on Xiaohongshu, a Chinese social platform for user-created lifestyle and product content. It can collect search results, candidate cover images, selected post details, and sanitized evidence.