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.
git clone --depth 1 https://github.com/nicelight/DevRails-26Wrote 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/commands/nicelight/devrails-26/grill-me)<a href="https://agentmods.dev/commands/nicelight/devrails-26/grill-me"><img src="https://agentmods.dev/badge/commands/nicelight/devrails-26/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 | $0.00015 | $0.00820 |
| Opus 5 | $0.00008 | $0.00410 |
| Sonnet 5 | $0.00003 | $0.00164 |
| Haiku 4.5 | $0.00002 | $0.00082 |
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 2d 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.
What it actually says
/grill-me - Product discovery by grilling
Read the existing brief and relevant project evidence when present. If only
legacy .memory-bank/analysis/product-brief.md or
.memory-bank/analysis/brainstorming/BR-*.md exists, use it as migration
context. Before interpreting product intent, sources, or operator answers,
load and apply the installed creator-vibe skill.
Work in rounds. The frontier is every decision whose prerequisites are settled: the questions that can be asked now without guessing. Ask the whole frontier in one round, numbering each question and giving a recommended answer. A question that depends on another open question belongs to a later round.
Format each question like this:
❓ **Q1** - **<question title>**: <question body>
➡️ <recommended answer>
Each answer reshapes the tree. Recompute the frontier before the next round.
Finding facts is the agent's job. Use repository evidence, tools, research, and bounded exploration as useful rather than asking the operator for facts that can be discovered. Running research delays only the dependent branches. Product decisions remain the operator's and require an explicit answer.
Use intent, users, problem, desired outcome, value, product direction, MVP boundary, non-goals, success, constraints, risks, assumptions, and open questions as a coverage lens, not a fixed questionnaire.
The session ends when the frontier is empty. Ask the operator to confirm that the understanding is shared before advancing.
<durable_output>
After each answered round, create or update
.memory-bank/analysis/index.md and .memory-bank/analysis/brief.md so the
session can resume without chat history. Keep material current findings,
sources, explicit decisions and rationale, constraints, rejected or deferred
directions, and open questions; do not preserve the interview transcript.
Use this minimum brief shape:
---
description: Durable Product Brief produced by /grill-me.
status: draft
type: product-brief
decision: blocked
---
# Product Brief
## Intent
## Problem / Users / Outcome
## Product Direction
## MVP Boundary / Non-goals
## Findings / Sources
## Decisions / Rationale
## Constraints / Risks / Assumptions
## Rejected / Deferred
## Open Questions
## Decision
blocked
Keep recommendations and unresolved choices separate from accepted decisions.
After the operator confirms shared understanding, set status: active,
decision: proceed, and the final ## Decision to proceed. Update the
analysis index with the brief link, decision, and immediate next command.
</durable_output>
<agent_discretion> Choose the exploration order, ideation techniques, tools, number of useful directions, and grouping of findings according to the maturity of the idea. Stop expanding when additional ideas no longer improve the operator's decision. Keep a clear concept short; explore a genuinely divergent problem more deeply. </agent_discretion>
<handoff_contract>
A confirmed brief with decision: proceed hands off to /write-prd when
project principles are already ratified|partial|skipped; otherwise it hands
off to /constitution. A blocked brief continues through /grill-me.
</handoff_contract>
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.
- 2d ago First seen · 103 lines · 15 tokens per session scan A cbb5b72cb541
grill-me is a command published in the GitHub repository nicelight/DevRails-26 (22 stars, last pushed 2d ago), licensed MIT. It adds 15 tokens to every session and 820 once invoked, about $0.0001 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 2026-09-05.
Other commands, from other repositories
template
Manage issue templates for streamlined issue creation.
sync-linear
Sync current work with Linear ticket status.
add-note
Add an internal or external note to a ConnectWise PSA ticket.
fest-show
Show festival progression (in-progress tasks, roadmap, and dependency view).
dispatcher
Pick the next-best repo to work on across the portfolio — rank free repos, recommend one, claim its lease atomically, and route to the entry command.
workpm
A project-management workflow for coordinating multiple AI workers through five stages. It includes task assignment, shared activity logs, worker replacement, and final checks.