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 skills/nicelight/devrails-26/mb-initnpx skills add nicelight/DevRails-26 --skill mb-initgit 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/skills/nicelight/devrails-26/mb-init)<a href="https://agentmods.dev/skills/nicelight/devrails-26/mb-init"><img src="https://agentmods.dev/badge/skills/nicelight/devrails-26/mb-init.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 | $0.00415 |
| Opus 5 | $0.00010 | $0.00208 |
| Sonnet 5 | $0.00004 | $0.00083 |
| Haiku 4.5 | $0.00002 | $0.00042 |
Grade A, and why
mb-init 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 4d 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
mb-init — External bootstrap router
This package does not create or sync Memory Bank files. Bootstrap is owned only
by scripts/install-framework.mjs in an available DevRails checkout.
Route
-
Verify the target repository path.
-
Use a DevRails checkout path only when the operator supplied it or the checkout and its installer are already verifiable. Never guess or invent the checkout path.
-
For a target without
.memory-bank/, return:node <devrails-checkout>/scripts/install-framework.mjs --bootstrap-only --target <target-repo> --yes -
For an explicit framework sync of an existing target, update runtime command skills and Memory Bank managed assets from the same prepared source copy:
node <devrails-checkout>/scripts/install-framework.mjs --bootstrap --sync --target <target-repo> --yes--bootstrap-only --syncrepairs only Memory Bank managed assets and does not update runtime command skills. Do not present it as a full framework sync.
Replace both placeholders with verified, shell-safe paths. If the checkout is
unknown or unavailable, ask the operator to provide one or run the external
installer action, then rerun /mb-init; never present a placeholder as an
executable command.
Do not run local bootstrap logic, copy a helper into the target, install a
dependency, or create the skeleton manually. When .memory-bank/ already
exists and no sync was requested, make no changes. Hand off to /start
only when its SKILL.md exists in the active .agents/skills/ or
.claude/skills/ runtime surface. Otherwise report that the skeleton is ready
and stop; do not claim /start is available. Runtime command installation
remains an explicit external installer action.
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.
- 4d ago First seen · 45 lines · 19 tokens per session scan A 8ffa08bcabc2
mb-init is a skill published in the GitHub repository nicelight/DevRails-26 (22 stars, last pushed 3d ago), licensed MIT. It adds 19 tokens to every session and 415 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 skills, from other repositories
best-practices
Searchable knowledge base of 152+ programming best practices across 30+ languages and frameworks. BM25-powered search over curated resources from industry leaders (Google, Airbnb, Uber, Mozilla, Shopify, OWASP).
rushdb-domain-template
Design a tailored RushDB schema for any use case through guided conversation. Use this skill when a user is starting a new project, unsure what labels to define, asking "what records should I create", or wants a scaffold for a known domain (e-commerce, healthcare, SaaS, CRM, agent memory, etc.). Interviews the user…
screen-repository
Screen an untrusted checkout for auto-executing hooks and unpinned dependencies before reading or running anything in it. Required first step of add-memory-system and reanalyze-memory-system. Use whenever a repository is cloned onto this machine for analysis.
platform-device-compatibility-matrix
Use when a game needs a platform and device compatibility matrix across OS, hardware, GPU, memory, resolution, input, network, storefront, certification, test evidence, and support policy; not for store metadata preparation.
native-skia
Use this skill for React Native Skia canvas animations, drawing primitives, Reanimated integration, CanvasKit/web caveats, performance, memory, and Expo/native validation. Trigger on React Native Skia, @shopify/react-native-skia, Skia Canvas, CanvasKit, Skia animation. Do not use for near-miss tasks outside these…
platform-guide
Answers how-to questions about third-party platforms (e.g. 'how to change theme in Shopify', 'set up a Stripe webhook', 'connect a domain in Vercel', Meta/TikTok/Google consoles, Laravel Forge, app stores) with fresh data: always WebSearch the current official guide first and cite it — never answer from stale memory.…