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 RBraga01/builder-growth --skill retention-designgit clone --depth 1 https://github.com/RBraga01/builder-growthWrote 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/rbraga01/builder-growth/retention-design)<a href="https://agentmods.dev/skills/rbraga01/builder-growth/retention-design"><img src="https://agentmods.dev/badge/skills/rbraga01/builder-growth/retention-design/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/rbraga01/builder-growth/retention-design"><img src="https://agentmods.dev/badge/skills/rbraga01/builder-growth/retention-design.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 | $0.00045 | $0.01792 |
| Opus 5 | $0.00023 | $0.00896 |
| Sonnet 5 | $0.00009 | $0.00358 |
| Haiku 4.5 | $0.00005 | $0.00179 |
Grade A, and why
retention-design 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 10d 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 — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Retention Design
The Law
A FEATURE LAUNCHED WITHOUT A DESIGNED RETENTION LOOP IS A FEATURE DESIGNED TO CHURN.
"Users will come back if they like it" has no activation moment, no habit trigger, and no reactivation path — which describes every churned product, because users who liked it also forgot about it.
Activation moment + habit loop + reactivation path defined before launch IS retention design.
When to Use
Trigger before:
- Launching any new product or significant new feature
- Redesigning an existing flow where D7 or D30 retention is below target
- Adding a feature whose primary value is realised over repeated sessions (not first use)
- Building any AI assistant, tool, or agent that requires habit formation
When NOT to Use
- One-time-use features where the user's goal is fully completed in a single session (e.g., a one-time document export — retention is not the right metric)
- Features for internal tools where users are required to return (no voluntary retention decision being made)
The Three Retention Elements
1 — Activation Moment
The specific action a user takes that predicts they will return.
Activation is not the first session. It is the moment in the first session when the user understands what the product can do for them — the "aha moment."
How to find it:
- Compare the behaviour of retained users (D30+) vs. churned users
- Find the action that is significantly overrepresented in retained users in the first session
- Verify causality vs. correlation: does encouraging new users to reach this action earlier improve D30?
What it looks like:
Slack: send your first message
Dropbox: put a file in a folder and access it from another device
GitHub Copilot: accept the first autocomplete suggestion
Required in the design:
- Name the activation action
- Measure how many new users reach it in the first session (baseline)
- Design the first-session flow to lead users to this action
2 — Habit Loop
The cue → routine → reward cycle that brings users back.
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.
- 10d ago First seen · 168 lines · 45 tokens per session scan A 0df73c7885fb
retention-design is a skill published in the GitHub repository RBraga01/builder-growth (2 stars, last pushed 2mo ago), licensed MIT. It adds 45 tokens to every session and 1,792 once invoked, about $0.0002 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
gr-geo-cite
A guide for checking whether AI assistants such as ChatGPT, Claude, Perplexity, and Gemini mention or cite your website when answering set questions.
gingiris-reddit-marketing
🇺🇸 Reddit Marketing & Growth Playbook 2026 — The complete Reddit operations SOP for AI/SaaS founders. Covers account warming (Karma 0→500 in 20 days), shadow ban prevention, multi-account matrix via fingerprint browsers, subreddit selection across 10K+ communities, content templates (90/10 rule), Product Hunt ×…
gingiris-github-star-growth
A guide to steadily increasing the number of people who bookmark an open-source project on GitHub by building content, community, events, contributors, and ambassadors.
gr-blog-post
A Jekyll blog publishing workflow. Jekyll is a tool that turns text and templates into a website, and this workflow also prepares matching English, Japanese, and Korean posts.
gr-seo-patrol
An automated SEO and GEO checking routine. SEO helps pages appear in search engines; GEO checks how they may be found and understood by AI-based search tools.
gingiris-growth-finder
Diagnose a product's growth model, stage, and current constraint, then route the request to the narrowest Gingiris specialist and execute it when installed. Use for broad or uncertain growth questions involving go-to-market, Product Hunt, GitHub stars, open-source marketing, B2B SaaS, PLG, ASO, SEO/GEO, AI citations…