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 instructions/deepsqlai/deepsql/claude-mdgit clone --depth 1 https://github.com/DeepSQLAI/deepsqlWhat 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 | $0.16976 | $0.16976 |
| Opus 5 | $0.08488 | $0.08488 |
| Sonnet 5 | $0.03395 | $0.03395 |
| Haiku 4.5 | $0.01698 | $0.01698 |
Grade B, and why
deepsql CLAUDE.md scanned grade B with 2 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 yesterday.
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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
`sudo -u postgres psql`, which only exists on a bare-metal install — on the Compose Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -X POST http://localhost:8080/api/users/admin/bootstrap \ How it starts
The opening of the file, as written. The whole thing — 922 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — DBA Agent
Full reference: See
docs/root/CLAUDE.mdfor comprehensive architecture, all API endpoints, entity details, and integration flows. Architecture summary: SeeAGENTS.mdfor high-level codebase map used by all AI agents.
Project Overview
DBA Agent is an AI-powered Database Performance Assistant with autonomous troubleshooting capabilities. Monorepo with a Java backend and a React frontend.
Tech Stack:
- Backend: Spring Boot 4.0.3 (Java 25), Spring AI 2.0.0-M2, PostgreSQL vault DB
- Frontend: React 19.2.3, Vite 7.3.0, Tailwind CSS 4.1.18
- AI: bring-your-own LLM via
LlmProviderRegistry— OpenAI, Azure OpenAI, or any OpenAI-compatible server. Vector store: pgvector or Azure AI Search (RAG) - Caching: Redis/Valkey
- Databases supported: PostgreSQL, MySQL (via provider registry pattern)
Development Commands
Backend (Spring Boot)
cd backend
mvn clean install # Build
mvn spring-boot:run # Run (dev mode, auth disabled)
mvn spring-boot:run -Dspring-boot.run.profiles=prod # Run (prod, auth enabled)
mvn test # Run all tests
mvn test -Dtest="*IntegrationTest" # Integration tests only
Backend URL: http://localhost:8080/api
Frontend (React + Vite)
npm install # Install dependencies
npm run dev # Dev server (http://localhost:3000)
npm run lint # Lint
npm run test:local-regression # Quick local regression suite
npm run local-deploy # Start services + run local regression suite
npm run mcp:phase1 # Start the Phase 1 DeepSQL MCP server (stdio)
npm run build # Build (dev)
npm run build:production # Build (prod)
Dev credentials: There is no baked-in admin/admin login — AuthController.login requires a
real User row matched by email, not username, so a fresh database (new Postgres volume)
has no account to log in with at all. SECURITY_AUTH_ENABLED=false only bypasses JWT/MCP token
validation (JwtAuthenticationFilter, McpTokenAuthenticationFilter); it does not create a
user or skip the login form. Create the first admin via the bootstrap endpoint, gated by
SECURITY_ADMIN_BOOTSTRAP_ENABLED=true + ADMIN_BOOTSTRAP_SECRET, and only callable from
localhost:
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.
- yesterday First seen · 922 lines · 16,976 tokens per session scan B d42d13a212ad
deepsql CLAUDE.md is an instructions file published in the GitHub repository DeepSQLAI/deepsql (23 stars, last pushed 2d ago), licensed Apache-2.0. It adds 16,976 tokens to every session, about $0.0849 per session on Opus 5. A static security scan graded it B with 2 findings (asks for root, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
coolify AGENTS.md
Instructions for coollabsio/coolify, covering agents.md, project overview, design reference, development environment and start dev environment (uses docker-compose.dev.yml).
whodb AGENTS.md
Instructions for clidey/whodb, covering whodb development guide, git commits — hard rule, terminology, non-negotiable rules and behavioral guidelines.
graphjin AGENTS.md
Instructions for dosco/graphjin, covering graphjin agent guide, architectural overview, directory structure & responsibilities, build commands and coding guidelines.
databasement CLAUDE.md
Instructions for David-Crty/databasement, covering claude.md, project overview, development commands, setup and installation and running the application.
whodb CLAUDE.md
Instructions for clidey/whodb, a project described as: Where data access meets operational intelligence.
cli AGENTS.md
Instructions for planetscale/cli, covering planetscale cli — agent guide, public repository safety, concepts, flag placement and correct.