Full-Stack AI Agent Template generates full-stack AI applications with a FastAPI backend and Next.js frontend, including agents, retrieval-augmented generation, streaming, authentication, and integrations. It is for building AI products with features such as chat, conversation sharing, administration, and multiple agent or vector-database choices. Catalogue add-ons support the generated applications and their agent workflows.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/vstorm-co/full-stack-ai-agent-templatenpx agentmods add skills/vstorm-co/full-stack-ai-agent-template/rag-knowledgeWrote 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/vstorm-co/full-stack-ai-agent-template/rag-knowledge)<a href="https://agentmods.dev/skills/vstorm-co/full-stack-ai-agent-template/rag-knowledge"><img src="https://agentmods.dev/badge/skills/vstorm-co/full-stack-ai-agent-template/rag-knowledge/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/vstorm-co/full-stack-ai-agent-template/rag-knowledge"><img src="https://agentmods.dev/badge/skills/vstorm-co/full-stack-ai-agent-template/rag-knowledge.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00076 | $0.00687 |
| Opus 5.5 | $0.00030 | $0.00275 |
| Sonnet 5.5 | $0.00015 | $0.00137 |
| Haiku 4.5 | $0.00008 | $0.00069 |
Grade A, and why
rag-knowledge 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
RAG Knowledge Base ({{ cookiecutter.vector_store }})
The RAG stack lives in backend/app/services/rag/ (ingestion, vectorstore, embeddings, connectors). Retrieval is exposed to the agent as the search_knowledge_base tool, and to operators via the CLI and the dashboard.
CLI (run from backend/)
uv run {{ cookiecutter.project_slug }} cmd rag-ingest ./docs/ --collection docs --recursive # ingest files/folder
uv run {{ cookiecutter.project_slug }} cmd rag-search "your question" --collection docs # semantic search
uv run {{ cookiecutter.project_slug }} cmd rag-collections # list collections
uv run {{ cookiecutter.project_slug }} cmd rag-stats # chunk/vector counts
uv run {{ cookiecutter.project_slug }} cmd rag-drop <collection> --yes # delete a collection
Ingestion = parse → chunk → embed → upsert into {{ cookiecutter.vector_store }}. Re-ingesting the same source updates it (use --no-replace / --sync-mode to control dedupe).
Sync sources (connectors)
Connectors keep a collection in sync with an external source (Google Drive, S3/MinIO) on a schedule, and can be managed per-organization from the dashboard (/orgs/[id]/integrations) or via CLI:
uv run {{ cookiecutter.project_slug }} cmd rag-sources # list
uv run {{ cookiecutter.project_slug }} cmd rag-source-add # add (interactive)
uv run {{ cookiecutter.project_slug }} cmd rag-source-sync --all # trigger a sync
Connector credentials are encrypted at rest with CHANNEL_ENCRYPTION_KEY (Fernet).
Adding a new connector type
Implement a connector in backend/app/services/rag/connectors/ following the existing Google Drive / S3 connectors, register it in the connector registry, and expose its config fields. See docs/howto/add-sync-connector.md and docs/howto/configure-sync-sources.md.
Tuning retrieval
- Chunk size/overlap and parser (PyMuPDF / LlamaParse) are configured via env — see
docs/configuration.mdanddocs/rag.md. - Reranking (Cohere or local CrossEncoder) improves result ordering when enabled.
- If search returns poor results: confirm the collection is populated (
rag-stats), check the active collection in the chat's KB selector, and verify the embedding provider/key.
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 · 49 lines · 76 tokens per session scan A 89d38d46356a
rag-knowledge is a skill published in the GitHub repository vstorm-co/full-stack-ai-agent-template (1,951 stars, last pushed yesterday), licensed MIT. It adds 76 tokens to every session and 687 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 2026-08-30.
Other skills, from other repositories
video_action_recognition
Technical guidance for recognizing human actions in video without training a new machine-learning model or using an LLM. It describes a pipeline that samples video frames, estimates body poses, measures movement, and assigns labels with rules.
docker_packaging
A set of rules and templates for packaging a Python MCP service in Docker. Docker packages an application and its dependencies into a repeatable container that can run on another machine.
ai-portable-setup
Erstellt einen portablen KI-Arbeitsbereich auf einem USB-Stick oder beliebigen Laufwerk. RAG-Pipeline mit lokalen LLM-Modellen (Ollama), Vektordatenbank (ChromaDB) und vorkonfigurierten Prompts.
document-chunker
Split documents into overlapping token chunks for RAG pipelines and LLM context windows. Zero dependencies.
knowledge-search
Answer questions from the organization's indexed documents using knowledgesearch, and cite every claim back to a retrieved passage. Use whenever a question is about document content. Covers the search-first rule, answering only from the closed set of returned passages, the Sources list format, and how to tell an empty…
RAG Reasoning
Visible reasoning contract for retrieval, evidence selection, and grounded synthesis.