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 skills/siddsachar/row-bot/developer_custom_toolsnpx skills add siddsachar/row-bot --skill developer_custom_toolsgit clone --depth 1 https://github.com/siddsachar/row-botWhat 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 | $0.00022 | $0.00445 |
| Opus 5 | $0.00011 | $0.00222 |
| Sonnet 5 | $0.00004 | $0.00089 |
| Haiku 4.5 | $0.00002 | $0.00044 |
Grade A, and why
developer_custom_tools 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.
What it actually says
Developer Custom Tools
Use this skill when the user asks to add a GitHub repo, local folder, or current workspace as a reusable Row-Bot Custom Tool.
The user can ask in natural language. Treat requests like "make this repo a tool", "add this GitHub repo as a Custom Tool", or "turn this folder into a tool" as Custom Tool creation requests.
Workflow:
- Confirm the source: repo URL, local folder, or current workspace. For repo URLs, ask for an explicit clone parent folder if one is not already provided.
- Start a draft with
custom_tool_builderactionstartwhen that utility is enabled. It performs a compact repo scan plus one AI proposal pass, then validates the proposed commands. - Explain the draft commands in plain language. Do not ask the user to hand-write internal config files.
- Refine the draft with action
refinewhen the user asks for changes, or actionupdatefor a direct command/metadata edit. - Public repo Custom Tools require explicit user enablement of public repo tools before creation. Do not silently enable that gate.
- Create only after the user accepts the proposed commands. Use action
create. - Test at least one command with action
testbefore enabling or promotion. - Enable Developer use with action
enablewhen the user wants it active in Developer. - Promote with action
promoteonly when the user wants the tool available in normal chat/plugin tools. - If removing a draft or Custom Tool, use action
deleteand preserve the source folder unless the user explicitly asks to delete files.
Be clear about scope: a Custom Tool starts as Developer-managed metadata around an existing repo/folder; promotion makes it appear in Settings -> Plugins and normal chat tools.
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 · 36 lines · 22 tokens per session scan A ac05c29542e8
developer_custom_tools is a skill published in the GitHub repository siddsachar/row-bot (1,479 stars, last pushed 4d ago), licensed Apache-2.0. It adds 22 tokens to every session and 445 once invoked, about $0.0001 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 skills, from other repositories
codex-universal-proxy
Install and manage the unofficial codex-universal-proxy local compatibility tool. Use when the user wants Codex to install it from npm, clone/download the repo, configure model routing, start/restart/inspect the proxy, switch the local experimental Ollama/proxy workflow on or off, or uninstall it.
hermit-audit
Hermit + Claude Code 종합 감사 (read-only) — 8축 28항목 체크리스트로 Hermit/CC 환경(토큰 경제, 모델 라우팅, 스킬·훅·룰·MCP·세션 습관)을 평가. Git-shipped audit tool. 2026 best-practice adoption 검증용. Trigger: "/hermit-audit", "hermit audit", "hermit 감사", "hermit 점검", "토큰 감사", "cache hit rate", "MCP 오버헤드", "토큰 최적화 점검".
architecture-discipline
Загружай первым: границы слоёв (адаптер→core→executor→tools/services), JSON-only цикл, local-first, sqlite-vec, без новой архитектуры без согласования; указатель на остальные скиллы.
skill-authoring
Как добавлять и менять скиллы/промпты ассистента в .agents/: формат SKILL.md, каталог scripts/, зеркала (.claude, AGENTS.md, README) и проверка checkagentssync.py.
skill-generator
Generates AND modifies complete skills from natural language descriptions. Guides the user through need analysis and archetype selection, produces a skill package compliant with the agentskills.io standard, and imports it directly into the user's skills. Also adjusts, enriches or fixes an existing user skill by…
opencode-bridge
Auto-trigger for OpenCode, Zen cloud models, Ollama local models, and free/open-source LLM requests. Routes to /opencode command.