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 codebygarv/Ai-skills --skill grill-megit clone --depth 1 https://github.com/codebygarv/Ai-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/codebygarv/ai-skills/grill-me)<a href="https://agentmods.dev/skills/codebygarv/ai-skills/grill-me"><img src="https://agentmods.dev/badge/skills/codebygarv/ai-skills/grill-me.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 5.1 | $0.00041 | $0.00503 |
| Opus 5 | $0.00020 | $0.00251 |
| Sonnet 5 | $0.00008 | $0.00101 |
| Haiku 4.5 | $0.00004 | $0.00050 |
Grade A, and why
grill-me 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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Purpose
Challenge the user's idea, implementation, architecture, or decision as hard as a skeptical, experienced peer would — instead of defaulting to agreement or polite hedging. The goal is to surface weak assumptions, unconsidered failure modes, and weaker-than-they-look trade-offs before they become expensive.
When to Use
- The user explicitly asks to have an idea, plan, or architecture "challenged," "grilled," "stress-tested," or "poked full of holes."
- Before committing to a significant technical or product decision.
- When a plan sounds too clean and hasn't been argued against yet.
Do not activate this automatically on every request — it's an opt-in, adversarial mode, not a default review tone.
What to Analyze
- Restate the idea in one sentence to confirm you understood it before attacking it.
- Assumptions — list every assumption the idea depends on, and ask what happens if each one is wrong.
- Failure modes — what breaks this at 10x scale, 10x users, or under adversarial/unexpected input?
- Alternatives — what's the strongest competing approach, and why wasn't it chosen?
- Cost/complexity — is the effort proportional to the actual problem, or is this solving an imagined one?
- Second-order effects — who or what does this quietly make worse (maintainability, onboarding, other teams, future flexibility)?
Output Format
- Lead with the single sharpest objection, not a warm-up.
- Organize remaining pushback as a numbered list, strongest point first.
- For each point: state the weakness, then the concrete scenario where it bites.
- Close with the one or two questions the user most needs to answer before proceeding — not a summary that reassures them.
Avoid
- Softening every criticism with a compliment sandwich — say the hard thing plainly.
- Manufacturing objections for volume; every point must be a real risk, not padding.
- Attacking the person instead of the idea.
- Ending on a "but overall this is great!" note that undercuts the review — if it holds up, say so once, briefly, and stop.
What ships with it
2 files 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.
- 4d ago First seen · 40 lines · 41 tokens per session scan A 60c58714caa6
grill-me is a skill published in the GitHub repository codebygarv/Ai-skills (25 stars, last pushed 19d ago), licensed MIT. It adds 41 tokens to every session and 503 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
bun-api
Bun runtime API reference for TypeScript scripts. Covers Bun.serve() HTTP/HTTP/2 server with routes and WebSockets, fetch() transport options, Bun.file(), Bun.write(), Bun.$() shell, Bun.spawn(), Bun.Glob, Bun.env, bun:sqlite, Bun.sql() for PostgreSQL/MySQL via DATABASEURL, Bun.s3 for S3-compatible storage, Bun.redis…
bun-cli
Bun CLI reference for package management, script running, testing, bundling, and compilation. Covers bun install/add/remove/update, bun run, bun test, bun build, bunx, bun patch, bun audit fix, bun dedupe, bun prune, bun why, bun pm, bun repl, bunfig.toml, bun.lock, workspace catalogs, isolated installs and the global…
deps-upgrade
Validate JavaScript/TypeScript dependency upgrades after an interactive or manual update. Establish the real delta from git, detect peer/engine/type conflicts, extract breaking changes, migrations and adoptable features from release notes, assess held-back or seemingly unused packages, and run verification gates.…
git-pr
PR and MR workflows for GitHub (gh) and GitLab (glab). Creation, review comment handling, thread resolution, review state queries, merging, cost-aware bot review rounds (GitHub: Copilot, CodeRabbit), and Copilot code review configuration (rulesets, custom instructions, billing). Use when creating PRs/MRs, addressing…
git-project
GitHub Projects (v2) setup and management via gh CLI + GraphQL: organize a repo's issues into epics with native sub-issues, drive a board Status flow (Todo -> In Progress -> Done), set Project Priority, classify with native issue Types (Task/Bug/Feature) instead of labels, scope work with milestones (incl. resolving…
git-ci
CI/CD status queries for GitHub Actions (gh) and GitLab CI (glab). Check pipeline status, failing jobs, workflow runs, job logs, and merge readiness. Use when checking CI status, debugging build failures, viewing job logs, watching pipeline progress, or configuring gh/glab CI read-only allowlists. Not for PR workflows…