discuss

discuss is a command for coding agents from nicelight/DevRails-26. It costs 12 tokens per session (797 once invoked), scanned A, original, MIT.

A clarification command for resolving unknown requirements or contradictions before planning or coding.

In plain words
What is it for?
Use it to identify the right project document for an unanswered question, discuss the issue, and record an accepted answer through that document's normal owner.
Why use it?
It prevents the agent from inventing product, design, task, or testing decisions when the project information is incomplete.

Command

Install

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.

agentmods
npx agentmods add commands/nicelight/devrails-26/discuss
Clone the repo
git clone --depth 1 https://github.com/nicelight/DevRails-26

Wrote 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.

agentmods badge for discuss

README.md
[![agentmods](https://agentmods.dev/badge/commands/nicelight/devrails-26/discuss.svg)](https://agentmods.dev/commands/nicelight/devrails-26/discuss)
Your own site
<a href="https://agentmods.dev/commands/nicelight/devrails-26/discuss"><img src="https://agentmods.dev/badge/commands/nicelight/devrails-26/discuss.svg" alt="Measured on agentmods" height="20"></a>
Per session 12 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 797 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00012 $0.00797
Opus 5 $0.00006 $0.00398
Sonnet 5 $0.00002 $0.00159
Haiku 4.5 $0.00001 $0.00080

Measured 3d ago against content hash bd592633a2fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

discuss 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.

skills/_shared/references/commands/discuss.md · 87 lines

What it actually says

/discuss - Clarify before work

<input_contract> Identify the discussion scope from $ARGUMENTS or ask the operator to name it. Read only the relevant PRD, Memory Bank artifacts, protocols, and accepted decisions needed to understand that scope. Respect their normal source-of-truth precedence and ownership.

Before asking a material question, identify an existing owning canonical artifact and section where an accepted answer can be applied. If that durable destination does not exist, stop and route the unresolved question to the immediate owning skill to create or update its normal artifact, then resume /discuss. /discuss must not create the owning artifact or a decision log. </input_contract>

<hard_invariants>

  • /discuss does not invent product, architecture, contract, task, tier, or verification decisions.
  • It does not create a new status model or bypass the owning skill's gate.
  • A recommendation is advisory until the operator explicitly accepts it.
  • Do not solicit or treat an answer as accepted until its existing owning canonical artifact is identified as the durable destination.
  • Keep unresolved dependent work blocked; independent read-only analysis may continue when it does not pre-decide the choice. </hard_invariants>

<operator_decisions> Identify ambiguities, decisions needed, and risks. Whenever a real branch can affect the current outcome or downstream contract, ask the operator. Adapt the format and grouping to the issue: one question, a small group of tightly related questions, multiple choice, or open form. Explain impact and optionally give a recommendation with rationale.

Do not use a fixed round size, questionnaire UI, or answer-length limit. Do not ask formal questions for issues already resolved by authoritative evidence. Silence or an unanswered recommendation is not a decision. </operator_decisions>

<required_outputs> Apply every accepted answer durably to an existing canonical artifact owned by the relevant stage, such as PRD, requirements, feature, spec, or protocol. Remove contradictory superseded wording and record unresolved questions with their affected scope and owner in that artifact.

Update a scoped .protocols/<ID>/decision-log.md only when it already exists. Do not create a decision log, owning canonical artifact, or another artifact/registry family.

Do not create a separate interview registry. </required_outputs>

<agent_discretion> Choose evidence order, tools, analysis depth, question grouping, and concise recommendation form. Use the smallest discussion that resolves the material branch and do not restate the entire downstream workflow. </agent_discretion>

<handoff_contract> Return to the immediate owning skill: /brief, /constitution, /write-prd, /spec-init, /prd-to-features, /feature-doctor FT-<NNN>, initial /spec-design, or post-backbone /spec-redesign according to the artifact and blocker being resolved. /discuss never routes directly past a required product/design gate or to execution without an already-ready indexed task.

The result names every changed artifact path and section/anchor when practical, gives a concise accepted-decision summary, and lists each unresolved question with its owner. Preserve the owning skill's existing next gate and immediate handoff. </handoff_contract>

Changes

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.

  1. 3d ago First seen · 87 lines · 12 tokens per session scan A bd592633a2fe

Subscribe to this mod's changes

discuss is a command published in the GitHub repository nicelight/DevRails-26 (22 stars, last pushed 6d ago), licensed MIT. It adds 12 tokens to every session and 797 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 2026-08-30.