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/u9401066/nsforge-mcp/agents-mdgit clone --depth 1 https://github.com/u9401066/nsforge-mcpWrote 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/u9401066/nsforge-mcp/agents-md)<a href="https://agentmods.dev/instructions/u9401066/nsforge-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/u9401066/nsforge-mcp/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 | $0.00836 | $0.00836 |
| Opus 5 | $0.00418 | $0.00418 |
| Sonnet 5 | $0.00167 | $0.00167 |
| Haiku 4.5 | $0.00084 | $0.00084 |
Grade A, and why
nsforge-mcp 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 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.
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 — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NSForge Agent Harness
Workspace instructions for autonomous agents (Copilot, Cline, Codex, OpenHands, Hermes, …) working in the Neurosymbolic Forge (NSForge) repository.
Goal
Help build and operate NSForge — an MCP server that turns concepts into
verifiable, traceable entities (symbols → derivations → algorithms). The AI
orchestrates; deterministic tools reify. See
docs/reification-ladder-direction.md for the north star.
North Star
Every symbol / equation / value / line of code in a final result must have a tool call as its "birth certificate" (provenance). The AI must not hand-derive any of them.
Success = the amount the AI computes by hand approaches zero.
Working Style
- Use Traditional Chinese unless the user asks otherwise.
- Prefer exact file paths, command output summaries, and verification results.
- Offload mechanical/deterministic steps (symbolic calculation, simplification, code generation) to tools — do not hand-derive formulas.
- When changing behavior, add a focused regression test under
tests/.
Ground Truth: one command
Before and after changes, verify against the green baseline:
python scripts/check.py # all gates
python scripts/check.py --json # machine-readable summary (for agents)
python scripts/check.py --gates lint,type,test # subset
Gates: 12 total — lint, format, type, import, manifest, mcp, test, bench, generic, provenance, harness, diff. Exit code 0 = green.
(bench = derivation-correctness of benchmarks/*.json through the L3 orchestrator;
generic = arbitrary unseen compositions derive correctly, proving NSForge is a
derivation calculus, not a hand-built formula library; provenance = every
benchmark derivation carries a complete tool-provenance ledger, no hand-derived leaks;
mcp = MCP 2.1.1 discovery, schemas, metadata, payloads, resources, prompts, and
legacy-client compatibility stay green;
harness = the harness self-checks its own invariants — package version parity,
self-describing tools, and gate/doc parity — so agents can trust the signal.)
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 · 77 lines · 836 tokens per session scan A 983d9dc62f09
nsforge-mcp AGENTS.md is an instructions file published in the GitHub repository u9401066/nsforge-mcp (4 stars, last pushed 4d ago), licensed Apache-2.0. It adds 836 tokens to every session, about $0.0042 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
vera CLAUDE.md
Claude Code instructions for aallan/vera, covering claude.md — project orientation for claude code, virtual environment, key commands, project layout and writing vera code.
vera AGENTS.md
AGENTS.md instructions for aallan/vera, covering agents.md — instructions for ai agents, for agents writing vera code, conformance programs as reference, workflow and commands.
sagemath-mcp CLAUDE.md
Instructions for XBP-Europe/sagemath-mcp, covering claude.md, project, commands, linting and testing.
sagemath-mcp AGENTS.md
Instructions for XBP-Europe/sagemath-mcp, covering agent playbook, environment & setup, fast commands, where code goes and testing expectations.
ai-dial-core CLAUDE.md
Claude Code instructions for epam/ai-dial-core, covering claude.md, build & run, set credentials via environment variables, build (skip tests) and run all tests.
ai AGENTS.md
AGENTS.md instructions for vercel/ai, covering agents.md, project overview, repository structure, key directories and core package dependencies.