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/metadist/synaplan/agents-mdgit clone --depth 1 https://github.com/metadist/synaplanWrote 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/instructions/metadist/synaplan/agents-md)<a href="https://agentmods.dev/instructions/metadist/synaplan/agents-md"><img src="https://agentmods.dev/badge/instructions/metadist/synaplan/agents-md.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 | $0.06005 | $0.06005 |
| Opus 5 | $0.03002 | $0.03002 |
| Sonnet 5 | $0.01201 | $0.01201 |
| Haiku 4.5 | $0.00600 | $0.00600 |
Grade C, and why
Synaplan scanned grade C 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 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
docker compose exec -T backend sh -c 'rm -rf var/cache/test && mkdir -p var/cache/test && chmod -R 777 var/cache/test' Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- Test endpoints via Swagger UI at `http://localhost:8000/api/doc` (not curl/Postman). Before creating a new endpoint, check if a similar one exists. How it starts
The opening of the file, as written. The whole thing — 352 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Synaplan Development Guide
Full-stack AI knowledge management platform: RAG with MariaDB VECTOR + Qdrant, embeddable chat widgets, WhatsApp/Email integration, multiple AI providers (Ollama, OpenAI, Anthropic, Groq, Gemini, xAI).
Stack: PHP 8.3/Symfony 7, Vue 3/TypeScript/Vite, Docker Compose, frankenphp/caddy.
Repository Architecture
| Repo | Purpose |
|---|---|
synaplan/ |
Main app source code, local development environment (includes Qdrant) |
synaplan-base-php/ |
Base Docker image (FrankenPHP + gRPC + whisper.cpp) |
synaplan-platform/ |
Production deployment configs, pulls pre-built images |
Local dev (synaplan/): Builds from source, includes all dev tools (phpMyAdmin, MailHog, Vite dev server). Qdrant runs as a Docker service for AI memories and RAG vector search.
Production (synaplan-platform/): Uses ghcr.io/metadist/synaplan:latest, multi-server with external Galera cluster. See _devextras/SYSADMIN-help.md.
Critical Rules
Language
- Code, comments, commit messages: ALWAYS English. Never write German (or any other language) in code or comments.
- Chat responses: in the language the user chooses.
Docker Environment
All backend/frontend tooling runs inside containers:
docker compose exec backend php bin/console cache:clear
docker compose exec -T frontend npm run check:types
Prefer the make targets (they wrap docker compose exec correctly).
The database is MariaDB, not MySQL — the db container has no mysql binary. Use the mariadb client with the app credentials from docker-compose.yml:
docker compose exec -T db mariadb -usynaplan_user -psynaplan_password synaplan -e "SELECT ..."
Git — allowed, but NEVER on main
- Git operations (branch, add, commit, push) are allowed.
- NEVER commit or push directly to
main; never force-pushmainormaster. All changes go through feature branches + PRs. - Use Conventional Commits:
feat:,fix:,refactor:,docs:,chore:,test:— e.g.feat(frontend): add runtime config API support. - NEVER add attribution ("Generated with Claude Code", "Co-Authored-By: …", or similar).
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 · 352 lines · 6,005 tokens per session scan C 95d9b313b360
Synaplan is an instructions file published in the GitHub repository metadist/synaplan (111 stars, last pushed 4d ago), licensed Apache-2.0. It adds 6,005 tokens to every session, about $0.0300 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, 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
MemOS CLAUDE.md
Claude Code instructions for MemTensor/MemOS, covering claude.md, claude code entry, sub-agents and project knowledge.
memsearch CLAUDE.md
Claude Code instructions for zilliztech/memsearch, covering claude.md, build & test commands, install in development mode, run a single test file and run a specific test.
MemOS AGENTS.md
AGENTS.md instructions for MemTensor/MemOS, covering agents.md, project overview, repository layout, command cheatsheet and core api.
rag-rat AGENTS.md
Instructions for cq27-dev/rag-rat, covering agents.md, prefer the rag-rat mcp for code browsing, record durable learnings as rag-rat memories, repo orientation and worktree correctness.
obsidian-notes-rag CLAUDE.md
Instructions for proofsh/obsidian-notes-rag, covering claude.md, project overview, commands, install dependencies and run tests.
llm-brain CLAUDE.md
Instructions for kimsanguine/llm-brain, covering llm-brain — claude code 운영 지침, 가드레일 (절대 위반 금지), 프라이버시 경계 (agent memory os), 명령어 and ingest.