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 agentmods add skills/yoeld-wix/quiet-bash/deterministic-firstnpx skills add yoeld-wix/quiet-bash --skill deterministic-firstgit clone --depth 1 https://github.com/yoeld-wix/quiet-bashWrote 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/yoeld-wix/quiet-bash/deterministic-first)<a href="https://agentmods.dev/skills/yoeld-wix/quiet-bash/deterministic-first"><img src="https://agentmods.dev/badge/skills/yoeld-wix/quiet-bash/deterministic-first.svg" alt="Measured on agentmods" 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 | $0.00058 | $0.00996 |
| Opus 5 | $0.00029 | $0.00498 |
| Sonnet 5 | $0.00012 | $0.00199 |
| Haiku 4.5 | $0.00006 | $0.00100 |
Grade A, and why
deterministic-first 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 4d 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 — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pipe for the Answer, Don't Read the Haystack
quiet-bash's hooks shrink output you already chose to read. This is the other
move: change the choice. A human debugging a 10k-line log doesn't read it into
their head — they pipe grep | sort | uniq -c. The shell is cheaper (no tokens
for the haystack, none to re-send it every later turn) and more correct (a
uniq -c count doesn't hallucinate; a regex doesn't skim past line 4,000).
The decision rule
Before issuing a Read / cat / large fetch whose purpose is to locate or
compute something, ask: can a pipeline return just the answer? If yes, run
the pipeline and let only the answer enter context.
| Task | Read-to-answer (avoid) | Deterministic form |
|---|---|---|
| Find where / which files | open files until you spot it | rg -l PAT · rg -n PAT · grep -rln PAT |
| Count / frequency | read log, tally by eye | grep -c PAT · quiet-agg FILE 'PAT' |
| Extract fields | copy values out of JSON/text | jq '.field' · awk '{print $2}' · grep -oE |
| Parse structured slices | scan a config visually | jq / yq / quiet-query FILE keys |
| Config value — one field from a config/lockfile | read the whole file to find it | quiet-conf FILE KEY (jq path for json/yaml, var name for .env) |
| Code archaeology — who/when/which-commit | scroll full git log/blame, read files |
quiet-hist PATH · quiet-hist --pick STR · quiet-blame FILE S E |
| Orient in an unfamiliar repo / env | ls/find/read files; probe node -v/which X |
quiet-map (sizes/churn/tree) · quiet-env (pkg-mgr, CLIs, versions) |
| Apply a diff — existing patch / multi-file / no Edit tool | re-emit whole files; reason if a patch fits | quiet-applies FILE (does it fit?) then quiet-patch FILE; single small edits → use Edit |
| Math / dates / ordering — compute in-head | sum/percent/date-diff/order by reasoning | awk/bc · date -d · tsort (topological order) |
| Verify a fact / triage a log | read output to confirm or count failures | quiet-verify FILE 'PAT' · quiet-check FILE (PASS/FAIL + error tally) · test -f |
| Repeat over a set | N near-identical tool calls | xargs · for f in …; do …; done |
| Wait for a condition | poll by re-reading status | quiet-wait 'COND' --timeout N · until COND; do sleep N; done |
| Re-read a file you already read | re-open it (re-bills the bytes) | don't — its contents are already above this turn; scroll up |
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.
- 4d ago First seen · 48 lines · 58 tokens per session scan A a57cc52877a4
deterministic-first is a skill published in the GitHub repository yoeld-wix/quiet-bash (5 stars, last pushed 1mo ago), licensed MIT. It adds 58 tokens to every session and 996 once invoked, about $0.0003 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
lithium-kb
Generates, maintains, and visualizes a structured Markdown knowledge base (.lithium-kb/{architecture,debug,tasks,features}) with an interactive Neural Network graph, auto-watch mode, custom agent directives (.agentrules), and MCP server integration for Pi, Claude, Codex, Cursor, and other coding agents. Triggers…
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
interview
Ask one useful structured question at a time only when material product/implementation choices are genuinely missing; remember answers and produce a brief/spec. Discoverable facts should be investigated instead of asked.
writing
将共享历史中的已验证事实和计算结果整理成符合受众、格式与长度约束的成稿。.
test
Detect the project’s test stack, run the narrowest useful tests, create tests when authorized, and report coverage/gaps honestly.
verify
Exercise the real app/API/CLI and collect observable evidence; tests alone do not count as end-to-end verification.