Scaffolds a new System 1 agent module for a named task in the system1-agents repo, after a fit probe, with its test and README row, verified model by model.
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agent --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/build-s1a-agent .claude/skills/build-s1a-agent && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .claude/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agentType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agent --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/build-s1a-agent .agents/skills/build-s1a-agent && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .agents/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agent --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/build-s1a-agent .cursor/skills/build-s1a-agent && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .cursor/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/ThinkFlowLab/system1-agents.git --path .claude/skills/build-s1a-agent--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agent --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/build-s1a-agent .gemini/skills/build-s1a-agent && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .gemini/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agentInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/build-s1a-agent .github/skills/build-s1a-agent && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .github/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install ThinkFlowLab/system1-agents build-s1a-agent --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/ThinkFlowLab/system1-agents.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/build-s1a-agent .opencode/skills/build-s1a-agent && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "build-s1a-agent" agent skill from https://github.com/ThinkFlowLab/system1-agents/tree/main/.claude/skills/build-s1a-agent into .opencode/skills/build-s1a-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "build-s1a-agent", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
build-s1a-agentScaffolds a new System 1 agent module for a named task in the system1-agents repo, after a fit probe, with its test and README row, verified model by model.
This skill is specific to the s1a repository, where an agent is one module under s1a/agents/ that ends in a frozen SPEC, while the loop, decision-model layer, rethink rail, job folders and CLI are shared. It starts with a one-message intake of five facts: the task and what a finished episode looks like, where state comes from, how options are listed at each step, the score and any baseline, and whether a step needs arithmetic, constraint deduction, search or generated text, which counts as a stop.
A fit gate comes before any code. You write 8 to 12 hand-made cases as JSONL and run uv run s1a probe, which reports fits at 80 percent or more over at least 8 cases, too few cases below 8, or not a decision-model task. A wrong case that needed deduction or arithmetic stops the work even if the total passes. The agent then chooses a front (tool, browser or rail), copies the matching template, and writes the module, a test and a README row. Reference files describe the fit rule, the three fronts and state design.
6 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 3a2c2c6. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
uvFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use uv, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
System 1 Agent Builder loads about 1.7k tokens when it runs, and up to ~3.8k if it reads all its reference files. Until then it costs about 118 tokens; SKILL.md has 811 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from ThinkFlowLab/system1-agents at commit 3a2c2c6, republished under its Apache-2.0 licence (© ThinkFlowLab). 811 words, ~1,655 tokens.
.claude/skills/build-s1a-agent/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.A System 1 agent is one module under s1a/agents/ that ends in a frozen SPEC. The loop, the decision-model layer, the
rethink rail, the job folders and the CLI are shared and never change for a new agent. This skill produces the
module, its test and a README row, and stops at the first gate that fails.
Read references/fit-rule.md before the intake and references/state-design.md before the scaffold. The three
fronts and every spec field are in references/fronts.md.
Ask for, or confirm from the request, these five facts. Stop until they are clear.
Write 8 to 12 hand-written cases as JSONL, one decision each, spread over the task's situations, and run them:
uv run s1a probe cases.jsonlEach line: {"state": {...}, "options": {"key": "description"}, "rules": "...", "accept": ["key"], "note": "..."}.
The verdict prints as fits at 80 percent right or above over at least 8 cases, too few cases under 8, otherwise
not a decision-model task. A wrong case that needed deduction or arithmetic is a stop even when the total passes. Report the
table to the user. On a stop, say which cases failed and why, name the nearest task that would fit, and end.
ToolAgentSpec.BrowserAgentSpec.RailSpec.Copy the front's template from s1a/agents/_templates/ to s1a/agents/<name>.py and replace every
part; the template's comments say what each part is. SPEC.name is the module name. Write
tests/test_agents_<name>.py with the two tool-front classes of tests/test_templates.py, TestNimEnv and
TestNimThroughTheLoop, rewritten for the new module: the env's reset, candidates and winning line, then the rule
and random models through series.play with loop.WORKSPACE and series.optional_chat_model patched as there.
For a browser agent, copy TestBrowserTemplateOffline: the spec reaches the faked subagent through
support.browse_offline. For a rail, copy TestRailTemplateOffline: precision and recall on five hand-labelled
records through a ScriptedModel(noul=[...]) from s1a.decision_models. Add one row to the agents table in README.md. Follow references/state-design.md for the
observation, the candidate keys, the rules text and the budget. Do not touch s1a/tool/,
s1a/browser/, s1a/rails.py or s1a/spec.py.
Run each command, read its output, fix the agent before the next rung. Stop at the first rung that fails.
uv run pytest tests/test_agents_<name>.py -q # the adapter contract, no keys
uv run s1a run <name> --model random --rethink off --episodes 3 # mechanics through the loop, no keys
uv run s1a run <name> --model rule --rethink off --episodes 3 # when a baseline exists
uv run s1a run <name> --model jev --rethink off --episodes 3 --log # keys: latency, invalid keys must be 0
uv run s1a run <name> --model llm --rethink off --episodes 3 # the same seeds with the chat model
uv run python -m evals.table evals/results # one row per model
uv run pytest tests -q # the whole suite stays greenEvery run prints one JSON object: the series summary with scored, the episodes that got a score, errors, the
count the model could not play, and job_dir. random and rule exist for the tool front only; laya
(Laya in process, after uv sync --extra laya) runs wherever jev does, and so does cua (Cua-S1 Nano, after
uv sync --extra cua) except on a rail. For a browser or rail agent the
key-free rung is the offline test from step 4. The paid rung follows. A browser agent runs one task per call and
needs the chat-model key and a Jev key: uv run s1a run <name> --model jev --goal "...". A rail runs its
labelled set and needs a Jev key: uv run s1a run <name> --labelled-set records.jsonl.
Report the table, then prepare the demo/evidence handoff below. Series of a hundred episodes cost money; ask the user before starting one.
For a reusable use case, add a recipe using the recipe template and add-agent-recipe skill, then update the recipe index. Follow CONTRIBUTING.md's video guide for the required video of the application/task, the System1-Agents decision-model agent and System1-Omni inference in the same run. Show actual input, selected/executed actions, engine identity and checked result. Choose an application from the guide's README candidates that exercises the new agent's actual behavior. Use a terminal recording for a text agent or rail; label scripted checks and recorded replays. Use a supported System1-Omni path; record missing support or integration as a review gap. Follow PR #35's worked example linked in the guide. Keep an important PR draft until its video is supplied or a maintainer accepts the documented exception. Prepare the PR's Demo / evidence section with the clip, reproduction command and linked run artifacts. Recording, model execution and uploading each require the authorization/resources already established for the task. If any are unavailable, report the gap and hand off the prepared instructions.
© ThinkFlowLab, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 3 other files (references) in .claude/skills/build-s1a-agent of ThinkFlowLab/system1-agents.
Open the folder on GitHubat commit 3a2c2c6
System 1 Agent Builder next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| System 1 Agent Builder this skillThinkFlowLab/system1-agents | 126 | — | ~1.7k | Automated safety check: Pass | Apache-2.0 | |
| Google Agents CLI Adk Codepifferologo/cloud-agents-cli | 129 | 1 repos | ~768 | Automated safety check: Pass | Apache-2.0 | |
| Building Multi Connector Agentairbytehq/airbyte-agent-sdk | 135 | — | ~1.7k | Automated safety check: Notes | Custom licence | |
| Google Agents CLI Workflowpifferologo/cloud-agents-cli | 129 | 1 repos | ~5.6k | Automated safety check: Notes | Apache-2.0 | |
| Build Dashclawucsandman/DashClaw | 310 | — | ~1.3k | Automated safety check: Pass | MIT | |
| Edgeone Makers ToolsTencentEdgeOne/edgeone-makers-tools | 1.9k | 1 repos | ~646 | Automated safety check: Pass | MIT |
pifferologo/cloud-agents-cli
This skill should be used when the user wants to "write agent code", "build an agent with ADK", "add a tool", "create a callback", "define an agent", "use state management", or needs ADK (Agent…
airbytehq/airbyte-agent-sdk
Builds a complete agent with multiple Airbyte connectors using PydanticAI or Claude SDK.
pifferologo/cloud-agents-cli
This skill should be used when the user wants to "develop an agent", "build an agent using ADK", "run the agent locally", "debug agent code", "test an agent", "deploy an agent", "publish an agent"…
ucsandman/DashClaw
Contribute to the DashClaw codebase — architecture, scaffolding, tests, CI
TencentEdgeOne/edgeone-makers-tools
EdgeOne Makers platform development router — the single entry point for building, storing data, and deploying on Tencent EdgeOne Makers.
aws/agent-toolkit-for-aws
A skill your agent uses when a developer wants to create a new agent project or get started with AgentCore.
ThinkFlowLab/system1-agents
Review pull requests for system1-agents with high-confidence, evidence-based feedback.
ThinkFlowLab/system1-agents
Delegates click-through web tasks, games and quizzes to S1A through its s1a command, or asks a fast decision model to pick one option from a list you provide.
ThinkFlowLab/system1-agents
Add or update runnable use-case recipes for existing System1-Agents agents, with setup, inference-backend configuration, independent result checks, demo evidence and troubleshooting.
ThinkFlowLab/system1-agents
Prepares a local self-review report for a System1-Agents change before a PR is opened, checking the full diff, tests, docs, artifacts and claims against evidence.
Categories
Scaffolds a new System 1 agent module for a named task in the system1-agents repo, after a fit probe, with its test and README row, verified model by model. This skill is specific to the s1a repository, where an agent is one module under s1a/agents/ that ends in a frozen SPEC, while the loop, decision-model layer, rethink rail, job folders and CLI are shared. It starts with a one-message intake of five facts: the task and what a finished episode looks like, where state comes from, how options are listed at each step, the score and any baseline, and whether a step needs arithmetic, constraint deduction, search or generated text, which counts as a stop.
System 1 Agent Builder fits situations like: adding a new System 1 agent for a game, task, site or rail; checking whether a task suits a decision model before writing code; choosing between the tool, browser and rail fronts for an agent.
Run `npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a claude-code`. Or copy the skill folder (.claude/skills/build-s1a-agent in ThinkFlowLab/system1-agents) into .claude/skills/build-s1a-agent in your project. Claude Code loads it when a task matches its description.
Run `npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a codex`. Or copy the skill folder (.claude/skills/build-s1a-agent in ThinkFlowLab/system1-agents) into .agents/skills/build-s1a-agent in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add ThinkFlowLab/system1-agents --skill build-s1a-agent -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/build-s1a-agent, .gemini/skills/build-s1a-agent, .github/skills/build-s1a-agent and .opencode/skills/build-s1a-agent in your project.
Going by SKILL.md and its folder, System 1 Agent Builder needs the command-line tools its instructions call (uv). Our summary lists: A checkout of the system1-agents repository; uv, to run the s1a CLI.
SKILL.md contains no URLs. Its commands use uv, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
System 1 Agent Builder is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 1.7k tokens (SKILL.md is roughly 6.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.1k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with System 1 Agent Builder: Google Agents CLI Adk Code (pifferologo/cloud-agents-cli, 129 stars), Building Multi Connector Agent (airbytehq/airbyte-agent-sdk, 135 stars), Google Agents CLI Workflow (pifferologo/cloud-agents-cli, 129 stars) and Build Dashclaw (ucsandman/DashClaw, 310 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
ThinkFlowLab (a GitHub organization) maintains it in ThinkFlowLab/system1-agents, which has 126 GitHub stars. The repository holds 5 skills in this directory. The repository was last updated on October 8, 2026.
Source: ThinkFlowLab/system1-agents on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.
