Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/yinanli1917-cloud/searching-apple-notesnpx agentmods add skills/yinanli1917-cloud/searching-apple-notes/grill-meWrote 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/yinanli1917-cloud/searching-apple-notes/grill-me)<a href="https://agentmods.dev/skills/yinanli1917-cloud/searching-apple-notes/grill-me"><img src="https://agentmods.dev/badge/skills/yinanli1917-cloud/searching-apple-notes/grill-me/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/yinanli1917-cloud/searching-apple-notes/grill-me"><img src="https://agentmods.dev/badge/skills/yinanli1917-cloud/searching-apple-notes/grill-me.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 | $0.00091 | $0.01114 |
| Opus 5 | $0.00046 | $0.00557 |
| Sonnet 5 | $0.00018 | $0.00223 |
| Haiku 4.5 | $0.00009 | $0.00111 |
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 12d 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 — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Brainstorm Engine (grill-me)
Interview me relentlessly about every aspect of this plan. Walk down each branch of the design tree, resolving dependencies one-by-one. For each question, provide your recommended answer.
Ask questions one at a time. If a question can be answered by exploring the codebase, explore the codebase instead of asking.
Why this skill writes files during the interview
If the brainstorm stays only in chat, a session timeout or /compact loses everything. Writing incrementally means every resolved question is immediately durable — the PRD builds up as you go, and a session crash loses at most the current question. The finalize step at the end cleans up the working draft into a polished document.
Setup
Before the first question, ensure a task exists:
# Check for active task
python3 scripts/codex_harness.py task current
# If none exists, create one
python3 scripts/codex_harness.py task create "<title from user's request>"
python3 scripts/codex_harness.py task start <slug>
Note the task slug — you'll use it for every command below.
The interview loop
Each turn of the brainstorm follows this pattern:
- Ask one question with your recommended answer
- Wait for the user's answer (they may accept your recommendation, modify it, or reject it)
- Immediately write the decision — this is the critical step, don't batch these
What "immediately write" means
After the user answers, run these before asking the next question:
# Step 1: Record the question (returns an ID like q-1)
python3 scripts/codex_harness.py task question ask <slug> "<your question>" --recommended-answer "<your recommendation>"
# Step 2: Record the user's answer (use the ID from step 1)
python3 scripts/codex_harness.py task question answer <slug> <id> "<their answer>" --decision "<one-line summary for prd.md>"
The --decision flag auto-appends to prd.md's Decisions section. Beyond that, also update the relevant section of prd.md directly:
- Scope/goal decisions →
## Goal - Feature requirements →
## Requirements - Implementation choices →
## Technical Approach - Quality bar decisions →
## Acceptance Criteria - Boundary decisions →
## Out of Scope
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.
- 12d ago First seen · 108 lines · 91 tokens per session scan A 7eebd80c5e6c
grill-me is a skill published in the GitHub repository yinanli1917-cloud/searching-apple-notes (2 stars, last pushed 1mo ago), licensed MIT. It adds 91 tokens to every session and 1,114 once invoked, about $0.0005 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
apple-notes
Use this skill when the user wants to interact with Apple Notes on macOS - creating, searching, reading, updating, deleting, organizing, or formatting notes and folders. This skill provides access to Apple Notes through MCP tools and includes safe formatting guidance.
notes-snapshot-control-room
This skill should be used when the user asks to "connect Apple Notes Snapshot to a host", "run notesctl mcp", "diagnose why Apple Notes Snapshot failed to attach", "separate AI Diagnose from MCP", or "verify the control-room proof path". It teaches local preflight, MCP wiring, capability boundaries, and proof-first…
aws-wechat-article-writing
A writing tool for Chinese WeChat public-account articles, creating a first draft from a topic or outline and revising existing text. It can use a configured language-model service, such as DeepSeek, OpenAI, or Anthropic.
aws-wechat-article-images
An image-generation tool for Chinese WeChat public-account articles, making covers and illustrations based on an article’s title and content. It can create several visual options and reuse configured style presets.
aws-wechat-article-main
A coordinator for running a complete Chinese WeChat public-account publishing workflow. It connects separate steps such as choosing a topic, writing, reviewing, formatting, adding images, and publishing.
aws-wechat-article-assets
A local library for storing business descriptions, product images, and reusable writing or design presets for WeChat articles. It organizes these materials by product so other article tools can reuse them.