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/laymom/ai-prd-generatorWrote 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/laymom/ai-prd-generator/pm)<a href="https://agentmods.dev/commands/laymom/ai-prd-generator/pm"><img src="https://agentmods.dev/badge/commands/laymom/ai-prd-generator/pm/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/commands/laymom/ai-prd-generator/pm"><img src="https://agentmods.dev/badge/commands/laymom/ai-prd-generator/pm.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 | $0.00075 | $0.06509 |
| Opus 5.5 | $0.00030 | $0.02604 |
| Sonnet 5.5 | $0.00015 | $0.01302 |
| Haiku 4.5 | $0.00007 | $0.00651 |
Grade A, and why
pm 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 8d 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 — 315 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/pm — AI Feature PRD Generator
For people planning a feature that uses an LLM or ML model: before it is handed to engineering, make sure the failure handling, human review, and evaluation criteria are decided. The tool produces a PRD focused on WHAT and WHY (problem, users, scope, non-goals, user stories, acceptance criteria) plus an AI appendix, and starts with a short list of the decisions that are still open. HOW (architecture, implementation planning, design) is handed off to later steps.
If the feature turns out not to use a model, say so plainly and continue as a normal PRD without the AI appendix.
Write questions and the PRD in the language the user writes in.
Core principles
- 4 gated stages — at the end of each stage, do not move to the next stage until the user explicitly approves (in staged mode; see "Two modes").
- Ask only what is missing — do not re-ask anything the user already gave (problem, target users, constraints, and so on). Ask only for missing core context, batched into one short message.
- Never press — if the user says "I don't know", "you decide", or "later", do not ask follow-ups on that point. Record it as an
AI assumptionor anOpen questionand move on. - Parallel subagents (optional) — if the environment supports subagents, run background research during intake and handoff. If it does not, skip this and do everything sequentially in the conversation.
- Evidence labels — every claim in the PRD carries one of the 7 labels below. Keep the assistant's guesses separate from facts (user answers, code, documents).
Evidence labels
Use these exact label names. Keep them in English even when the PRD is written in another language (a translation in parentheses is fine), so labels survive handoff between tools.
| Label | Use for |
|---|---|
User-provided fact |
Something the user stated about their situation, data, or constraints |
Sourced fact |
A fact from a document, code, or web page — always include the link or file path |
Real user research |
Observations, quotes, interviews, reviews, logs, or data from real people |
AI assumption |
Anything the assistant inferred, proposed, or filled in without confirmation |
Synthetic persona response |
Output from simulated personas (for example a synthetic focus group). Never evidence; it goes under "Hypotheses to validate" |
Approved decision |
A specific choice the user explicitly approved |
Open question |
Something not yet decided or not yet known |
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.
- 8d ago Changed · +28 lines · +10 tokens per session db424a25c929
- 12d ago Changed · +60 lines · +7 tokens per session 9add0f6a6f32
- 1mo ago First seen · 227 lines · 58 tokens per session scan A 9a997ed8784b
pm is a command published in the GitHub repository laymom/ai-prd-generator (1 stars, last pushed 8d ago), licensed MIT. It adds 75 tokens to every session and 6,509 once invoked, about $0.0003 per session on Opus 5.5. 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 commands, from other repositories
recipe-blueprint
Define the structural design context needed by a prototype or PRD.
recipe-define
Create a PRD from validated product decisions and evidence.
recipe-discover
Frame product opportunities and create decision-relevant hypotheses.
recipe-persona
Create or update evidence-backed personas for product decisions.
recipe-prototype-prompt
Export an evidence-grounded prompt for an external prototype generator.
recipe-refine-visuals
Refine the auto-derived Visual Tokens with a design-capable human.