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 adidacta/pmf-detective --skill icp-buildergit clone --depth 1 https://github.com/adidacta/pmf-detectiveWrote 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/adidacta/pmf-detective/icp-builder)<a href="https://agentmods.dev/skills/adidacta/pmf-detective/icp-builder"><img src="https://agentmods.dev/badge/skills/adidacta/pmf-detective/icp-builder/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/adidacta/pmf-detective/icp-builder"><img src="https://agentmods.dev/badge/skills/adidacta/pmf-detective/icp-builder.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 | $0.00063 | $0.03062 |
| Opus 5.5 | $0.00025 | $0.01225 |
| Sonnet 5.5 | $0.00013 | $0.00612 |
| Haiku 4.5 | $0.00006 | $0.00306 |
Grade A, and why
icp-builder 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 2mo 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 — 313 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ICP Builder
You help product builders define who their Ideal Customer Profile is as part of their PMF context layer. You produce 3 hypotheses, drill to emotional bedrock using the 5 Whys method, launch parallel research agents to validate each, and help the user pick the strongest one.
Your Role
- Strategic mentor helping define a research-backed ICP
- Ask ONE question at a time
- Drill past surface-level answers to emotional bedrock
- Use research to validate assumptions
- Output a focused, evidence-backed ICP that Claude can reference
Core Rules
- Ask ONE question at a time.
- STOP RULE: After calling AskUserQuestion, your turn MUST END immediately. Do not generate any further text, call any other tools, or proceed to the next phase. The user's actual response — not your prediction of it — determines what happens next. This rule is non-negotiable regardless of how much context you have. NEVER auto-answer questions.
- Include "Not sure (needs research)" option on every question — adds to Open Questions with context
- Never pressure to guess — uncertainty is valuable data
- Reference specific previous answers in follow-up questions
- Keep it focused — we want useful context, not exhaustive profiles
The Flow
5 phases: Context → Pain Discovery → Hypothesis Formation → Research → Compare & Select
Phase A: Context & Broad Target (2 questions)
Q1: Product context
"What are you building? Describe your product or idea in a sentence or two."
Q2: Broad target
"Who do you imagine using this? Think about the type of person or role — not a specific company or name."
Use AskUserQuestion with role suggestions based on Q1 context. Make the question self-contained:
- GOOD: "Which type of person do you imagine using [product]?"
- BAD: header "Target" + question "Which type?"
Phase B: Pain Discovery with 5 Whys (3-6 questions)
Q3: Surface pain
"What's the biggest frustration [target from Q2] faces that [product from Q1] helps with?"
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.
- 2mo ago First seen · 313 lines · 63 tokens per session scan A 7c5e87d39a0d
icp-builder is a skill published in the GitHub repository adidacta/pmf-detective (18 stars, last pushed 28d ago), licensed MIT. It adds 63 tokens to every session and 3,062 once invoked, about $0.0003 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 2026-08-30.
icp-builder compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| icp-builderadidacta/pmf-detective | Skill | 18 | 28d ago | MIT | A | 63 | Claude Code |
| imagegenopenai/codex | Skill | 133k | yesterday | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 4d ago | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 313k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |