Borrowing it
Nothing to install: this file belongs to ironbee-ai/ironbee-devtools-skills. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ironbee-ai/ironbee-devtools-skills/main/AGENTS.mdgit clone --depth 1 https://github.com/ironbee-ai/ironbee-devtools-skillsWrote 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/ironbee-ai/ironbee-devtools-skills/agents-md)<a href="https://agentmods.dev/instructions/ironbee-ai/ironbee-devtools-skills/agents-md"><img src="https://agentmods.dev/badge/instructions/ironbee-ai/ironbee-devtools-skills/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 5.1 | $0.07008 | $0.07008 |
| Opus 5 | $0.03504 | $0.03504 |
| Sonnet 5 | $0.01402 | $0.01402 |
| Haiku 4.5 | $0.00701 | $0.00701 |
Grade A, and why
ironbee-devtools-skills 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 8d 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 — 281 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Guidelines for ironbee-devtools-skills
Keep this file up to date. When you change conventions, structure, or the relationship to the upstream
ironbee-devtoolsproject, update the relevant sections here as part of the same change. This file is the source of truth for AI coding agents working on this skills repository.
What this project is
ironbee-devtools-skills ships skills.sh-installable skills for IronBee DevTools. Each skill is a Markdown file (plus a references/ folder for CLI skills) that AI coding agents — Claude Code, Cursor, Windsurf, etc. — load to drive the IronBee DevTools CLIs.
This repo is purely documentation / agent prompts. There is no source code, no build step, and no runtime. Every change here is content edits to Markdown files, and correctness is judged by:
- Whether the CLI examples actually run against the IronBee DevTools binaries.
- Whether the documented flag names and input shapes match the tool schemas in the upstream
ironbee-devtoolsproject. - Whether the cross-references between SKILL.md files and
references/*.mdfiles resolve.
Upstream source of truth: ironbee-devtools
This repo is downstream of ironbee-devtools. Every CLI subcommand, every flag, every default value, every domain folder, every MCP server-instructions paragraph — all of that is generated by code in the upstream repo. If you change anything here, verify it against the upstream tool definitions, never the other way around.
Specifically:
- A tool's CLI subcommand name comes from its
name()method:domain_command-name→ CLIdomain command-name. Tools whose name has no underscore (e.g.scenario-add,execute) fall under the auto-generateddefaultdomain — with one special case:executeis special-cased torun executein the CLI runner.scenario-runis constructed with aToolRegistry(so scripts can compose other scenarios) and is therefore not in any platform'scliProvider.tools; the CLI runner appends it to every platform viawithSharedCliTools(src/cli/utils.ts), so it IS a direct CLI subcommand on all six platforms —cli default scenario-run --name '...'(flags--name,--args, repeatable--param key=value,--timeout-ms, plus the step-engine flags--resume,--resume-output,--cache,--step-by-step,--refresh,--persist-cache; no--scope). The scenario CRUD set (scenario-{add,update,delete,list,search,recorded-actions}) is registered on all six platforms too — each platform'stools/index.tsspreads the sharedtools/scenario(src/tools/scenario). - A tool's allowed CLI flags come from its Zod
inputSchema(): every camelCase field becomes a kebab-case flag. Object / record fields become a single JSON-valued flag (--metadata '{...}') — never split into per-key flags. Array fields render as a variadic flag (--flag <value...>) with no per-element JSON parsing, so each token is a raw string: arrays of scalars are passed as repeated/space-separated tokens (e.g. scenario-run's--param key=value --param key2=value2), and arrays whose elements are themselves objects (e.g. scenario-add's--params) effectively cannot be set from the CLI — declare them via MCP /run execute. Boolean fields typedz.boolean().optional()(without.default(true)) become a single switch — passing it sendstrue, omitting it sendsundefinedand the server applies its default.--no-<flag>is NOT auto-registered; onlyz.boolean().default(true)wraps generate both--flagand--no-flag. To forcefalsefrom CLI on these fields, call viarun execute/ MCP. - A tool's description in the skill prose should match the substance of
description()in code, not invent new behavior. - The CLI binary names live in upstream
package.json'sbinblock:ironbee-devtools-cli(browser default, aliasironbee-browser-devtools-cli),ironbee-node-devtools-cli,ironbee-python-devtools-cli,ironbee-backend-devtools-cli,ironbee-android-devtools-cli,ironbee-terminal-devtools-cli. Never use the legacybrowser-devtools-cli/node-devtools-cli/python-devtools-cli/backend-devtools-cli/android-devtools-cli/terminal-devtools-cli/browser-devtools-mcpnames. (Upstream also has aniosplatform folder, but it ships no CLI binary yet — do not add an iOS skill untilpackage.jsonbinlistsironbee-ios-devtools-cli.) - The
composemeta-platform is MCP-only — it has NO CLI skill. UpstreamPLATFORM=compose(COMPOSE_PLATFORMS=browser,backend,node,…) multiplexes several platforms' tools behind one MCP server, each under a prefix (browserbdt_, backendbedt_, nodendt_, pythonpdt_, androidadt_, terminaltdt_) — and since 0.25.0 the shared, platform-agnostic tools (execute,scenario-*,jira_*,linear_*,o11y_*) carry the common prefixibdt_instead of being exposed unprefixed (config.COMPOSE_COMMON_PREFIX). It shipsironbee-compose-devtools-mcpinbinbut noironbee-compose-devtools-cli(PlatformInfo.cliInfois omitted), so — exactly like iOS — do not add aironbee-compose-devtools-cli/skill until upstreampackage.jsonbinlists a compose CLI binary. The six CLI skills already document every tool compose can expose; the prefixing is an MCP-session concern, not a CLI one. - The default daemon port is
2020for all six CLIs. When multiple daemons run simultaneously the examples should pick different--portvalues (browser 2020, node 2021, backend 2022, android 2023, terminal 2024, python 2025 by convention).
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.
- 8d ago First seen · 281 lines · 7,008 tokens per session scan A e9b91488368b
ironbee-devtools-skills AGENTS.md is an instructions file published in the GitHub repository ironbee-ai/ironbee-devtools-skills (3 stars, last pushed 1mo ago), licensed MIT. It adds 7,008 tokens to every session, about $0.0350 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
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
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.
vscode buildNext.instructions.md
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).
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).
langchain AGENTS.md
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.
spec-kit AGENTS.md
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.