AIBAST Agents Library is a collection of industry-focused AI agent templates accompanied by a local server that connects agents to GitHub Copilot for language-model inference. It helps developers create and run tool-using agents and isolated project environments, with an optional cloud-backed path for persistent memory. The catalogue entries provide the repository's agents, skills, commands, hooks, and instructions.
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 skills add microsoft/aibast-agents-library --skill rapp-brainstemgit clone --depth 1 https://github.com/microsoft/aibast-agents-libraryWrote 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/microsoft/aibast-agents-library/rapp-brainstem)<a href="https://agentmods.dev/skills/microsoft/aibast-agents-library/rapp-brainstem"><img src="https://agentmods.dev/badge/skills/microsoft/aibast-agents-library/rapp-brainstem/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/skills/microsoft/aibast-agents-library/rapp-brainstem"><img src="https://agentmods.dev/badge/skills/microsoft/aibast-agents-library/rapp-brainstem.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 | $0.00026 | $0.01747 |
| Opus 5.5 | $0.00010 | $0.00699 |
| Sonnet 5.5 | $0.00005 | $0.00349 |
| Haiku 4.5 | $0.00003 | $0.00175 |
Grade A, and why
rapp-brainstem 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 1mo 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 — 142 lines — stays where its author put it; the contents beside it link to each section on GitHub.
RAPP Brainstem
The Brainstem is the product. Keep setup plumbing invisible.
When the user says "Give me my Brainstem," asks to install RAPP, or asks you to operate or maintain their Brainstem, own the complete local operation. Do not send terminal instructions back to the user. The only acceptable handoffs are unavoidable GitHub sign-in, operating-system elevation, tenant authorization, and explicit approval for a destructive action.
GitHub Copilot is the golden path. Claude Code follows this same contract as a
compatibility path. Set the lifecycle actor to github-copilot when running in
Copilot and claude-code when running in Claude Code.
Authority
- Treat this installed marketplace plugin as the lifecycle trust anchor.
Resolve the plugin root from this skill file on every request and execute
its bundled
rapp_operator/rappctl.pydirectly. Use only the adjacentscripts/bootstrap.*andinstaller-lock.json. - Do not download an operator, bootstrap script, or replacement lock from a
public manifest.
rapp-operator.jsonis informational, not bootstrap authority. - Never execute
~/.rapp/operator/rapp_operator/rappctl.pyor promote a marketplace bundle into a stable copied operator. Marketplace updates take effect because every lifecycle use resolves the currently installed plugin. - Treat the upstream
brainstem.pyand existing installer as Grail. Never edit either during setup, operation, repair, or maintenance. - Use
POST /chatas the only Brainstem execution entry point. Do not add or call a competing management route.
First decision: existing state or fresh setup
- Check whether
~/.brainstemexists before deciding to bootstrap. - If it exists, this is an existing install even when health is offline. Never
run fresh bootstrap over it. Run the current plugin's bundled
rappctl.py status, then follow Existing installs below. - Only when
~/.brainstemis absent, probehttp://127.0.0.1:7071/health/public. If another Brainstem is reachable, stop rather than adopting it. - If both state and health are absent, run the bootstrap bundled in this local
plugin:
- macOS/Linux:
scripts/bootstrap.sh --actor <actor> - Windows:
scripts/bootstrap.ps1 -Actor <actor>You execute it; the user does not.
- macOS/Linux:
- Do not pre-create a Python plan. The bootstrap works before Python exists. It
verifies the local plugin bundle, persists a pre-mutation envelope binding
the actor, installer-lock digest, exact installer URL and hash, target
rolling tag plus commit/tree/version, and operator hashes, then invokes the
unchanged installer with
--no-launch --version. - The bootstrap lock binds PID plus process creation identity. Any failure
after mutation restores the prior absent
~/.brainstemstate by moving the exact bootstrap-owned partial tree into private quarantine (or removing only that exact tree if quarantine cannot be activated). It never deletes pre-existing state. - The newly installed Brainstem Python runs
rappctl.pydirectly from the current plugin to reconcile the exact envelope. Stop on any mismatch or partial result. - A successful reconcile proves installation only. It deliberately records
pending-live-canaryand does not append live-verification evidence. - Use the Brainstem-managed Python for runtime start and verification:
- macOS/Linux:
~/.brainstem/venv/bin/python - Windows:
~/.brainstem/venv/Scripts/python.exe
- macOS/Linux:
- From the current plugin operator, run
plan start --actor <actor>, read the returnedplan_hash, then apply that exact plan withapply <plan_hash> --approve <plan_hash>. - The start apply proves public health and deliberately leaves live verification pending so first-run authentication can complete.
- Open the local Brainstem sign-in experience when authentication is required and let the user complete GitHub device authorization.
- After authentication, create and apply a separate
verifyplan. It must send a real canary throughPOST /chatusing{"user_input":"Confirm my Brainstem is alive and name one capability."}. - Finish only when
/chatreturns a real response and the separate verification frame exists. Report the RAPPID, Brainstem version, verified evidence head, and the result.
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.
- 1mo ago First seen · 142 lines · 26 tokens per session scan A 1e8bfbdb8ac5
rapp-brainstem is a skill published in the GitHub repository microsoft/aibast-agents-library (8 stars, last pushed yesterday), licensed MIT. It adds 26 tokens to every session and 1,747 once invoked, about $0.0001 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 2026-09-15.
Other skills, from other repositories
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…