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 titusz/skills --skill high-stakesgit clone --depth 1 https://github.com/titusz/skillsWrote 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/titusz/skills/high-stakes)<a href="https://agentmods.dev/skills/titusz/skills/high-stakes"><img src="https://agentmods.dev/badge/skills/titusz/skills/high-stakes/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/titusz/skills/high-stakes"><img src="https://agentmods.dev/badge/skills/titusz/skills/high-stakes.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 | $0.00173 | $0.02907 |
| Opus 5 | $0.00086 | $0.01453 |
| Sonnet 5 | $0.00035 | $0.00581 |
| Haiku 4.5 | $0.00017 | $0.00291 |
Grade A, and why
high-stakes 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 3d 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 — 178 lines — stays where its author put it; the contents beside it link to each section on GitHub.
High-Stakes
The task to execute is in $ARGUMENTS. Treat that text as the work to deliver, not as a description to discuss.
If $ARGUMENTS is empty (a bare /high-stakes, or "do this carefully" said about work already in the
conversation), the task is the most recent concrete request in the conversation. Name it explicitly
in your plan and confirm it before executing - never run the gauntlet against an empty task.
The organizing principle is watch like a hawk: a smart, hasty result is the default output of any single agent, and its mistakes are no longer typos - they are subtle conceptual errors a sloppy junior would make. The executor that wrote the artifact is its worst judge, so nothing it produces is trusted until independent skeptics try to break it and fail. Your job is to define what "correct and clean" means for this task, then loop that adversarial gauntlet until it comes back empty - agents do not tire, so the looping against a fixed target is the leverage.
The ten failure modes you are defending against
These are the specific bugs the skeptics hunt for. Every verifier maps to one. The plan-first gate (step 2) targets modes 1–6, where a guess becomes invisible the moment you write confident prose over it; execution and the gauntlet (steps 3–4) target modes 7–10.
- silent-assumptions - filled an underspecified detail (path, data shape, edge case, version, default) with a guess and built on it without stating it. The most common failure.
- unmanaged-confusion - hit genuine uncertainty and picked the first reading instead of asking.
- hidden-inconsistencies - found a contradiction (spec vs. code, conflicting requirements) and silently picked a side instead of surfacing it.
- unsurfaced-tradeoffs - several viable approaches existed; committed to one as if inevitable.
- no-pushback - implemented a flawed premise faithfully instead of challenging it.
- sycophancy - reflexive agreement and praise standing in for honest judgment.
- subtle-conceptual-errors - compiles/reads clean but is wrong: off-by-one, bad boundary, misused API semantics, race, broken invariant.
- overcomplication - 1000 lines where 100 would do; layers and knobs with one caller.
- dead-code - leftover scaffolding, unused symbols, commented-out blocks, dead branches.
- orthogonal-damage - touched, reformatted, or deleted code/comments unrelated to the task.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 3d ago Changed 53e51bc4b811
- 9d ago First seen · 178 lines · 173 tokens per session scan A a660e24f7196
high-stakes is a skill published in the GitHub repository titusz/skills (2 stars, last pushed 3d ago), licensed MIT. It adds 173 tokens to every session and 2,907 once invoked, about $0.0009 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
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…