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/frostney/gocciascript/agents-mdgit clone --depth 1 https://github.com/frostney/GocciaScriptWhat 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 | $0.01341 | $0.01341 |
| Opus 5 | $0.00671 | $0.00671 |
| Sonnet 5 | $0.00268 | $0.00268 |
| Haiku 4.5 | $0.00134 | $0.00134 |
Grade A, and why
GocciaScript 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 2d 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Instructions for AI coding assistants (Cursor, Claude Code, and similar) working in this repository.
Contributing guide vs this file
| Purpose | |
|---|---|
| CONTRIBUTING.md | Contributing guide for all contributors—humans and agents. Workflow, critical rules, testing, Pascal code style, formatting, build/run reference, documentation index. If it affects what may be merged, it belongs there. |
This file (AGENTS.md / CLAUDE.md) |
Agent-only context: how assistants should operate in this repo, where to read first, and what not to duplicate. It does not replace CONTRIBUTING. |
Assistants should treat CONTRIBUTING as authoritative for contribution requirements. Use this file for assistant-specific expectations and navigation, not a second copy of CONTRIBUTING.
Expectations for assistants
- Read CONTRIBUTING.md before substantive edits—especially Critical rules and Code style.
- Run verification yourself when the environment allows (tests, format check); do not only tell the human what to run unless execution is impossible.
- Match the project's workflow: branch from
main, focused diffs, tests and docs updated per CONTRIBUTING. - Treat project-local skills as external playbooks: files under
.agents/skills/are not normal repo documentation. Do not edit them unless the user explicitly asks to change that skill. Put repo-specific assistant expectations in this file, or create/update a separate skill only when explicitly requested. - Infer architecture boundaries during planning: when a change touches website routes, API handlers, generated reports, external services, credentials, artifacts, caches, CI outputs, or deployment/build steps, identify where the work belongs (build time, request time, client time, CI/scheduled time) and compare against existing project patterns before implementing. Do not rely on the user or a skill checklist to spell this out.
- Search the real source layout: when prompts, automations, or audits search this repo, target
source/units,source/shared,source/app,tests,scripts, andwebsite/srcexplicitly. Do not assume a generic rootsrc/tree;source/generatedis generated data and should only be inspected or regenerated when the task specifically requires it. - Do source deep-dives before policy claims: when a question asks how a runtime, standard, engine, or dependency behaves, treat README text, docs, comments, and prior notes as leads—not proof. Check the normative source first when there is one (for ECMAScript, ECMA-262/ECMA-402), then inspect the actual implementation paths in this repo and in any comparison engines or libraries named in the question. Record the specific clauses, files, gates/flags, and code paths that support the conclusion before recommending policy, scope, or architecture. If the answer depends on whether behavior is shim-level, parser-level, runtime-level, or object-model-level, classify each surface by the mechanism it actually needs.
- Clean first for stale FPC failures: after a merge, branch switch, PR sync,
generated resource change, or unexplained compiler/resource error, retry with
./build.pas --clean <target>(or./build.pas --clean) before diagnosing the reported source line. See Tooling — Stale FPC Build Artifacts. - Do not paste large chunks of CONTRIBUTING into this file when CONTRIBUTING changes—edit CONTRIBUTING instead, and keep AGENTS short.
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.
- 2d ago First seen · 67 lines · 1,341 tokens per session scan A 528ba8375535
GocciaScript AGENTS.md is an instructions file published in the GitHub repository frostney/GocciaScript (20 stars, last pushed 6d ago), licensed MIT. It adds 1,341 tokens to every session, about $0.0067 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 2026-08-30.
Other instructions, from other repositories
pascal-mcp-sdk AGENTS.md
Instructions for frostney/pascal-mcp-sdk, covering agent instructions, hard constraints, runtime / commands, code organization and testing.
tc39-atlas AGENTS.md
AGENTS.md instructions for bosens-China/tc39-atlas, a project described as: 面向中文用户与 AI Agent 的 TC39 提案索引、双语审查与 ECMAScript 现代化 Skill。.
pascal-mcp-sdk CLAUDE.md
Instructions for frostney/pascal-mcp-sdk, a project described as: FreePascal-native MCP (Model Context Protocol) server library — expose tools, resources, and prompts from Pascal programs to AI agents. Zero third-party dependencies, stdio + Streamable HTTP.
appwrite AGENTS.md
AGENTS.md instructions for appwrite/appwrite, covering appwrite, commands, stack, layout and libraries.
qiankun AGENTS.md
Instructions for umijs/qiankun, covering qiankun project knowledge base, structure, architecture (big picture), commands and build (vite 8 → package javascript; tsc → declarations).
joint CLAUDE.md
Instructions for clientIO/joint, covering claude.md, project overview, common commands, install dependencies and build distribution files (runs grunt).