# a5c-ai/babysitter

Babysitter enforces obedience on agentic workforces and enables them to manage extremely complex tasks and workflows through deterministic, hallucination-free self-orchestration

Repository: https://github.com/a5c-ai/babysitter
Canonical: https://ross.abutalabs.com/products/babysitter
Homepage: https://a5c.ai
Language: JavaScript
License: MIT
License Family: permissive
Topics: agent-orchestration, agent-skills, agentic-ai, agentic-workflow, ai-automation, babysitter, claude-code, claude-skills, trustworthy-ai, vibe-coding, ai-agents, claude-code-skills, codex-plugin, codex-skills, claude-code-workflows, claude-workflows, codex-workflow, hermes-plugin, pi-extension
Last push: 2026-08-26T07:08:59+00:00

## Health v2 (maintenance only)
Score: 75/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 78, longevity 17
- inputs: {"age_days": 240, "days_push": 7, "days_rel": 68, "gap_med": 0, "n_releases_24m": 110}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1738, forks 103 (observed 2026-08-28T04:05:29.969653+00:00)

## What it is
Babysitter is an npm-distributed CLI and runtime that enforces deterministic, step-by-step process control on AI coding agents across 12 supported harnesses (e.g. Claude Code, Codex). It defines workflows as code with quality gates, human approval breakpoints, and an immutable decision journal so agents cannot skip steps or hallucinate progress.

## Use cases
- enforce step-by-step discipline on AI coding agents
- run complex multi-step workflows with quality gates and human approvals
- prevent agents from drifting or hallucinating task completion
- orchestrate agent work across Claude Code, Codex, and other harnesses
- keep an immutable audit journal of agent decisions
- define reusable workflow blueprints for agentic coding tasks

## When to choose
- you use AI coding agents and need guaranteed process compliance
- you want deterministic, auditable agent workflows with gates and approvals
- you need harness-agnostic orchestration across multiple AI coding tools

## When to avoid
- you need a general-purpose multi-agent conversation framework rather than enforced coding workflows
- your project does not use Node.js 20+ or any supported AI coding harness
- you want fully autonomous agents without checkpoints or human approval

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, workflow-automation, developer-tools, cli, llm-inference
- domain: developer-tools, large-language-models
- platform: cli, cross-platform
- tags: agent-orchestration, claude-code, agentic-workflow, process-enforcement, hallucination-prevention, quality-gates, harness-agnostic, blueprints, npm-package, ai-agents, automation, nodejs

## Member repositories
- a5c-ai/babysitter (main) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:29.969653+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:30:35.406882+00:00, confidence not recorded.
  - readme: https://github.com/a5c-ai/babysitter (fetched 2026-08-28T04:05:29.969653+00:00, sha f85184202f47)
  - homepage: https://a5c.ai (fetched 2026-08-29T11:07:59.050367+00:00, sha 7145d1b8569d)
  - site_page: https://a5c.ai/company (fetched 2026-08-29T11:07:59.059626+00:00, sha 78b01fe92f8c)
- Data as of 2026-08-30T08:39:29.467469+00:00.