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/spec-autogit clone --depth 1 https://github.com/nicelight/DevRails-26What 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 | $0.00014 | $0.01374 |
| Opus 5 | $0.00007 | $0.00687 |
| Sonnet 5 | $0.00003 | $0.00275 |
| Haiku 4.5 | $0.00001 | $0.00137 |
Grade A, and why
spec-auto 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.
How it starts
The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/spec-auto - Autonomous SDD design
Supported arguments:
--init;- exactly one
FT-<NNN>.
<input_contract>
--init requires the normal /spec-init inputs after /write-prd and before
/prd-to-features.
FT-<NNN> requires:
- clarified PRD, product, requirements, epics, and product features;
.memory-bank/spec-backbone.mdand pure.memory-bank/spec-index.md;- a Global Backbone Status of
complete, or validminimalwith explicitnot_applicablerationale, plus positive integerPlanning Revision; if missing or invalid, route to/spec-design; .memory-bank/foundation.mdwith an explicit decision and, when foundation work is required, a closed Foundation Gate before product task handoff;- relevant canonical specs and accepted operator decisions;
.memory-bank/workflows/sdd-design-contract.md.
If no argument is given, infer --init only when clearly before
/prd-to-features; otherwise halt and request one explicit product feature.
Reject FT-000 and every multi-feature target.
</input_contract>
<hard_invariants>
/spec-autodoes not ask the operator during the unattended run and never treats that as permission to choose for them.--initinherits/spec-initsource authority and decision boundaries; it does not invent contracts, architecture, security, migration, irreversible data behavior, or Foundation sufficiency.- Apply the shared SDD authority and operator-decision contract in
FT-<NNN>mode; unresolved material choices block instead of receiving agent defaults. - In post-PRD modes, inherit the accepted global architecture and material module/slice map. Do not reopen, weaken, or replace them with a framework recommendation, feature-local heuristic, or one-slice-per-feature mapping.
- Preserve feature status vocabulary
spec_design_status: complete|not_required|blocked. - Do not create a new status or terminal-state model.
/spec-autopreserves Global BackbonePlanning Revision. Accepted backbone changes route to/spec-redesign; feature design never owns revision.- Current-state drift or baseline correction alone does not change Planning Revision while the accepted target remains unchanged. </hard_invariants>
<operator_decisions> Apply the shared operator-decision boundary. When it leaves a relevant material branch unresolved:
- write the exact question, alternatives/impact, affected artifacts/features, and decision owner to the existing backbone, feature, spec, or autonomous decision log as appropriate;
- set
Pre-PRD Spec Status: blocked, Global Backbone Statusblocked, or featurespec_design_status: blockedaccording to ownership; - stop with existing
HALT_BLOCKING_QUESTIONSorHALT_CLARIFICATION_REQUIRED; - report the interactive resume skill (
/write-prd,/spec-init,/feature-doctor,/spec-design, or/spec-redesign).
Non-decision implementation detail remains agent discretion. </operator_decisions>
<required_outputs>
Across modes, edit canonical artifacts only when durable framing/design state, routes, decisions, blockers, registry links, or applicable contracts have changed, or the existing contract needs repair. Do not append per-run audit/reconciliation history or mirror the final chat handoff.
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.
- 3d ago First seen · 149 lines · 14 tokens per session scan A f1b1d790fe80
spec-auto is a command published in the GitHub repository nicelight/DevRails-26 (22 stars, last pushed 5d ago), licensed MIT. It adds 14 tokens to every session and 1,374 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
git
Git operations with intelligent commit messages and workflow optimization.
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.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
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.