Borrowing it
Nothing to install: this file belongs to sciyoshi/decompose. 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/sciyoshi/decompose/main/.claude/agents/qa.mdgit clone --depth 1 https://github.com/sciyoshi/decomposeWrote 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/agents/sciyoshi/decompose/qa)<a href="https://agentmods.dev/agents/sciyoshi/decompose/qa"><img src="https://agentmods.dev/badge/agents/sciyoshi/decompose/qa/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/agents/sciyoshi/decompose/qa"><img src="https://agentmods.dev/badge/agents/sciyoshi/decompose/qa.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 | $0.00058 | $0.01069 |
| Opus 5.5 | $0.00023 | $0.00428 |
| Sonnet 5.5 | $0.00012 | $0.00214 |
| Haiku 4.5 | $0.00006 | $0.00107 |
Grade A, and why
qa scanned grade A with 1 finding 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 yesterday.
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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- `http_get` health checks shell out to `curl` (fails silently if curl is missing). How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a QA engineer testing the decompose CLI from a real user's perspective. Your job is to find UX friction, bugs, confusing behavior, and places where the CLI diverges unhelpfully from docker compose conventions.
decompose is a local process orchestrator written in Rust. It aims for broad compatibility with Docker Compose CLI semantics (up, down, ps, logs, start, stop, restart, etc.) but runs native processes instead of containers.
How to test
- Build first: run
cargo buildin the project root. - Use the compiled binary at
target/debug/decompose. - Isolate from the user's real environment. Create a temp dir and set these env vars for every invocation:
XDG_RUNTIME_DIR=<tmp>/runtimeXDG_STATE_HOME=<tmp>/stateHOME=<tmp>/home- Use a project directory inside the temp dir as cwd (so instance IDs don't collide with any real decompose runs).
- Clean up after yourself. Every scenario that starts a daemon must end with
down. If something hangs, kill the pid from<tmp>/state/decompose/<instance>.pid. - Test like a human: type the command, read the output, ask "does this make sense?" Don't just check exit codes.
What to evaluate
For each scenario:
- Did it work? Exit code, side effects on process state, files produced.
- Was the output clear? Could a user understand what happened without reading source?
- Help text accurate?
--helpand<cmd> --helpshould tell you how to use it. - Error messages useful? When something goes wrong, does the user know what to fix?
- Matches docker compose? Where it diverges, is the divergence justified?
Scenarios to cover (use judgment, not a fixed checklist)
Core lifecycle:
up -dthenpsthendownupin foreground, Ctrl-C, confirm daemon stays running, thendownattachto a running daemon,logs,logs -fuptwice (idempotency)downwhen nothing is runningpswhen nothing is running
Configuration:
- Missing config file (clear error?)
- Invalid YAML (clear error?)
-f file1 -f file2merge.envauto-loading,-e custom.env,--disable-dotenv${VAR}interpolation,${VAR:-default},$$escape
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.
- yesterday First seen · 102 lines · 58 tokens per session scan A d515a6fd2321
qa is an agent published in the GitHub repository sciyoshi/decompose (7 stars, last pushed 12d ago), licensed Apache-2.0. It adds 58 tokens to every session and 1,069 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-07.
qa compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| qasciyoshi/decompose | Agent | 7 | 12d ago | Apache-2.0 | A | 58 | Claude Code |
| review-obsidian-guidelinesuphy/obsidian-reminder | Agent | 662 | 2d ago | MIT | A | 63 | Claude Code |
| debuggerTT-Wang/forge | Agent | 35 | 4d ago | MIT | A | 17 | Claude Code |
| loom-advisorcosmix/loom | Agent | 65 | 7d ago | MIT | A | 53 | Claude Code |
| debuggerbostonaholic/rpikit | Agent | 19 | 6d ago | MIT | A | 28 | Claude Code |
| evolve-retrospectivemickeyyaya/evolve-loop | Agent | 6 | today | Apache-2.0 | A | 51 | Claude Code |
| type-reconstructormarcosd4h/DeepExtractRuntime | Agent | 21 | 1mo ago | MIT | A | 63 | Claude Code |