Happier is an open-source, end-to-end encrypted client that lets people run AI coding-agent sessions on a computer and continue controlling them from a phone, browser, or desktop app. It is for developers who use agents such as Claude Code, Codex, Gemini, or OpenCode across multiple devices, with self-hosting available. The catalogue entries are add-ons for configuring or extending this client and its supported coding-agent workflows.
Borrowing it
Nothing to install: this file belongs to happier-dev/happier. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/happier-dev/happier/dev/.agents/skills/decompose-gates/SKILL.mdgit clone --depth 1 https://github.com/happier-dev/happierWrote 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/happier-dev/happier/decompose-gates)<a href="https://agentmods.dev/skills/happier-dev/happier/decompose-gates"><img src="https://agentmods.dev/badge/skills/happier-dev/happier/decompose-gates/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/happier-dev/happier/decompose-gates"><img src="https://agentmods.dev/badge/skills/happier-dev/happier/decompose-gates.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00058 | $0.01153 |
| Opus 5.5 | $0.00023 | $0.00461 |
| Sonnet 5.5 | $0.00012 | $0.00231 |
| Haiku 4.5 | $0.00006 | $0.00115 |
Grade A, and why
decompose-gates 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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Decompose With Gates
Turn a hard task into pieces that can each pass or fail on their own, ordered so the riskiest assumption is tested first. Full doctrine: docs/agent-craft.md §2–§3.
Procedure
- Split along verification boundaries, not implementation convenience. Each piece must have its own pass/fail check that does not depend on the other pieces being right. If checking B assumes A is correct, A+B is one piece, not two.
- State each piece as a falsifiable claim, not a task. "The watermark advances only after server ack" (checkable) — not "update watermark logic" (a task). Attach to each claim the concrete check that decides it: a test slice, a log observation, a live replay, a measurement.
- Name the risk spots before sequencing. Write the two or three "if I'm wrong anywhere, it's here" spots — silent failure modes, irreversible steps, boring mechanical stretches — and design a specific verification for each. Generic suite runs are uniform effort against non-uniform risk.
- Order by information yield. Test first the smallest shared assumption whose failure would invalidate repeated downstream units. This gates only dependent replication: keep independent work moving and expand the rolling window as soon as the assumption holds. Skip the ramp when prior evidence or a deterministic tool already proves the unit shape.
- Relay dependency context, not only order. State the upstream contract and evidence each piece consumes. Those assumptions are themselves claims to check; most integration failures live at the interfaces.
- Right-size. A piece owns a whole responsibility end to end: analysis, change, tests, validation. Too small to fail meaningfully is overhead — micro-slicing (one-boolean extractions) provably grew the god-files it was meant to shrink. Too big to check independently is not decomposed yet.
- Prefer consumed verticals over horizontal spines. The first implementation slice should connect a real entry point through its canonical owner to an observable outcome. Do not split protocol/schema/producer/consumer into weeks of dormant layers whose correctness can only be known after activation. When compatibility forces phases, each phase names its active reader/consumer and removal condition.
- Firewall live-path corrections. If gated or dormant work exposes a live dispatch, lifecycle, persistence, migration, or startup defect, handle that correction as its own complete vertical at the live owner. Do not partially activate the dormant design to fix the production path.
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 c925c077e036
- 1mo ago First seen · 40 lines · 58 tokens per session scan A a9e303df8088
decompose-gates is a skill published in the GitHub repository happier-dev/happier (2,072 stars, last pushed yesterday), licensed MIT. It adds 58 tokens to every session and 1,153 once invoked, about $0.0002 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.
Other skills, from other repositories
agtx-oneshot
One-shot a whole project on an agtx board: decompose the goal, run every task unattended, unblock the workers, and merge each one. Use when the user wants a long autonomous run rather than a single session.
agtx-sweep
Sweep this conversation into agtx tasks and push them to the kanban board. Use when the user wants to capture, decompose, or hand off conversation results to the agtx board.
agtx-execute
Execute an approved implementation plan. Implement the changes, then write a summary to .agtx/execute.md and stop.
agtx-plan
Plan a task implementation. Analyze the codebase, create a detailed plan, write it to .agtx/plan.md, then stop and wait for user approval before making any changes.
agtx-research
Explore the codebase to understand a task before planning. Write findings to .agtx/research.md and stop. This is a read-only exploration — do not modify any files.
agtx-review
Self-review completed work. Check for correctness, edge cases, and code quality. Write review to .agtx/review.md and stop.