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/bitdg/tablerag/agents-mdgit clone --depth 1 https://github.com/BitDG/TableRAGWhat 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 | $0.00529 | $0.00529 |
| Opus 5 | $0.00264 | $0.00264 |
| Sonnet 5 | $0.00106 | $0.00106 |
| Haiku 4.5 | $0.00053 | $0.00053 |
Grade A, and why
TableRAG AGENTS.md 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 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.
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
TableRAG MCP Agent Notes
- Source discovery is non-recursive by default. Do not index subfolders unless the project YAML
explicitly sets
recursive: true; root-only discovery and a 413-file real build verified this boundary. - Keep the Python MCP SDK below major version 2 until the FastMCP server and protocol tests are
deliberately migrated. Allowing an unconstrained
mcp>=1.10installed an incompatible 2.0 API;mcp>=1.10,<2.0passed in-memory, real stdio, and Streamable HTTP calls. - On Windows, replace the local Web server by resolving the listening PID with
netstatand verifying the localhost page is TableRAG before stopping it. CIM andGet-NetTCPConnectionprocess discovery hung on this host; the signature-checked replacement restarted port 8766 and the diagnostics status endpoint returned HTTP 200. - On Windows, active stdio MCP clients can lock
.venv\Scripts\table-rag.exe, causinguv runto fail while refreshing the editable install. Do not stop user MCP services for tests; run.venv\Scripts\python.exe -m pytestdirectly and useuv lockfor lockfile-only updates. This path passed the full suite while four live MCP client processes remained running. - Do not write large DuckDB index tables with Python
executemany; use bounded vectorized bulk inserts and verify catalog counts. Replacing row-wise field-value writes produced 1,868,984 values from 413 real workbooks in 129.36 seconds, followed by a 1.54-second unchanged rebuild. - DuckDB connections in one process must use the same
read_onlyconfiguration. Web queries that overlap the index watcher use read-write connections with schema initialization disabled; the writer-overlap regression and 40 live/api/staterequests passed without HTTP 400 responses. - Repository checkouts may use the GitHub repository name as their folder. Tests for workspace
identity must compare resolved repository roots, not a hard-coded checkout basename; the
Windows/Linux CI matrix passed after the rename-safe assertion replaced
table-rag-mcp. - Bare mirror clones set
remote.*.mirror=true, which rejects explicit refspecs and--force-with-lease. For leased history rewrites, add a separate non-mirror remote; the two-branch atomic cleanup push passed with that setup.
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 · 29 lines · 529 tokens per session scan A 6e3631a28581
TableRAG AGENTS.md is an instructions file published in the GitHub repository BitDG/TableRAG (1 stars, last pushed 8d ago), licensed Apache-2.0. It adds 529 tokens to every session, about $0.0026 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 2026-08-31.
Other instructions, from other repositories
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
buildNext
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
spec-kit AGENTS.md
Instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
langchain AGENTS.md
Instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.