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 agentmods add commands/nicelight/devrails-26/prd-to-featuresgit 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/prd-to-features)<a href="https://agentmods.dev/commands/nicelight/devrails-26/prd-to-features"><img src="https://agentmods.dev/badge/commands/nicelight/devrails-26/prd-to-features.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 | $0.00019 | $0.01855 |
| Opus 5 | $0.00010 | $0.00928 |
| Sonnet 5 | $0.00004 | $0.00371 |
| Haiku 4.5 | $0.00002 | $0.00186 |
Grade A, and why
prd-to-features 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 5d 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 — 178 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/prd-to-features - Clarified PRD -> product plan
This command owns product decomposition, not PRD clarification, architecture design, testing policy, task generation, or implementation planning.
<input_contract> Require:
.memory-bank/prd.mdwithtype: prd,clarification_status: complete, andconstitution_checked: true;.memory-bank/spec-backbone.mdwithPre-PRD Spec Status: ready_for_prdand explicit decomposition inputs;- a pure
.memory-bank/spec-index.md.
Read the Constitution before writing. Use spec-backbone decomposition inputs and only canonical specs routed by spec-index that are relevant to the current PRD. Do not load every SDD spec by default.
Stop and route to the owning command when an input is missing, stale, placeholder-only, blocked, internally inconsistent, or has broken relevant links:
- PRD/Constitution issue ->
/write-prdor/constitution; - framing/index issue ->
/spec-init. </input_contract>
<hard_invariants>
- Do not write/clarify the PRD, perform architecture design, create TASK records
or implementation plans, choose test levels, define global testing gates, or
create/modify
.memory-bank/testing/*./prd-to-featuresdoes not create or modify files under.memory-bank/testing/. - Do not let a lower-precedence PRD or feature override Constitution or an authoritative canonical spec.
FT-000is reserved for Foundation Dev Path and is never a product feature.- Do not set feature
spec_design_status: completebefore the global design backbone exists and feature concerns are actually satisfied. - Do not add clarification metadata to every feature; it is only for an explicit feature-level decomposition blocker.
- Product feature acceptance criteria use stable
FT-<NNN>-AC-<NNN>IDs. Preserve an ID while its accepted behavior is the same; never renumber or reuse an existing ID. </hard_invariants>
<operator_decisions> The agent may choose ordinary decomposition tactics, but a real ambiguity or branch that could materially change product scope, REQ/EP/FT boundaries, acceptance, actors, behavior, ownership, or traceability belongs to the operator.
In interactive mode, ask one question or a small group of tightly related
questions, explain what the decomposition choice changes, and optionally
recommend an option. The recommendation is not accepted without an explicit
answer. Record accepted decomposition-only decisions in
.protocols/PRD-BOOTSTRAP/decision-log.md and the owning product artifact,
remove contradictions, then revalidate before deriving dependent docs. If the
answer changes PRD scope, behavior, acceptance, or another PRD-owned contract,
route it through /write-prd and stop this run before continuing decomposition.
In autonomous mode, use only an already accepted authoritative decision. If
none exists, record the question/affected artifacts and set terminal state
HALT_BLOCKING_QUESTIONS; do not write the affected decomposition as fact.
</operator_decisions>
<required_outputs>
Create .protocols/PRD-BOOTSTRAP/plan.md and decision-log.md if missing.
Update product.md with product identity, core value, audience, primary flow,
constraints, and non-goals. Update requirements.md with stable REQ-* IDs,
out-of-scope items, and RTM REQ -> Epic -> Feature -> Test.
Each epic records value, success metrics, acceptance criteria, evidence-backed
source/normative/constraint links when useful. Set epic status: draft by
default; promote it to document status: active only after its open questions
are closed. This document status is not the task lifecycle.
Each feature records:
- use cases, edge/failure behavior, and
## Acceptance Criteriawith one### FT-<NNN>-AC-<NNN>block per accepted criterion. Each block names at least one governingREQ-*on an exact- REQ: REQ-<NNN>[, ...]line, the observable criterion, and its verification method; - acceptance closure for every material product outcome: each observable
edge/failure outcome that can itself block acceptance or realize a
significant accepted risk names one or more covering AC IDs, or records an
authoritative
Disposition: out_of_scopewith its source andChange route: /write-prd. Several related outcomes may share one AC; - for each material NFR, an owning
REQ-*that preserves the observable quality, accepted numeric target or qualitative criterion, and general pass/fail-changing conditions. Its feature AC preserves the feature-specific application, observable criterion, and verification method; do not duplicate the full requirement across artifacts; - optional evidence-backed source artifacts, normative inputs, constraints/invariants, and verification targets;
- optional
## Behavior specsrouting only; do not create behavior JSON; status: draftby default;- an
## SDD Design Gatenote routing immediately to mandatory/spec-design, with feature-level canonical design completed later by/feature-to-tasksor/spec-autoafter any required Foundation Gate; - candidate canonical links only when grounded in evidence; omit
spec_design_statusunless its existing truthful state must be preserved.
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.
- 5d ago First seen · 178 lines · 19 tokens per session scan A 41b42325c656
prd-to-features is a command published in the GitHub repository nicelight/DevRails-26 (22 stars, last pushed today), licensed MIT. It adds 19 tokens to every session and 1,855 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.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.