Babysitter is a workflow engine for AI coding agents that enforces predefined steps, quality checks, human approvals, and decision records. It is used to coordinate complex, repeatable agent workflows across supported coding tools. The catalogue contains skills, agents, instructions, settings, a plugin, and an MCP integration for its workflow.
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/a5c-ai/babysitter/knowledge-curationnpx skills add a5c-ai/babysitter --skill knowledge-curationgit clone --depth 1 https://github.com/a5c-ai/babysitterWrote 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/skills/a5c-ai/babysitter/knowledge-curation)<a href="https://agentmods.dev/skills/a5c-ai/babysitter/knowledge-curation"><img src="https://agentmods.dev/badge/skills/a5c-ai/babysitter/knowledge-curation.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 | $0.00033 | $0.00379 |
| Opus 5 | $0.00016 | $0.00189 |
| Sonnet 5 | $0.00007 | $0.00076 |
| Haiku 4.5 | $0.00003 | $0.00038 |
Grade A, and why
knowledge-curation 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 yesterday.
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
- After completing work, BEFORE PR creation (reflect mode)
- When recovering from context loss (recovery priming)
Knowledge Categories
| Category | File | Content |
|---|---|---|
| Critical Rules | facts.jsonl | MUST FOLLOW constraints |
| Gotchas | gotchas.jsonl | Common pitfalls |
| Patterns | patterns.jsonl | Codebase best practices |
| Decisions | decisions.jsonl | Architectural choices with rationale |
| Anti-Patterns | anti-patterns.jsonl | What NOT to do |
| Codebase Facts | codebase-facts.jsonl | Structural information |
| API Behaviors | api-behaviors.jsonl | Undocumented quirks |
Process
Prime Mode
- Load knowledge base files for work type
- Surface MUST FOLLOW rules first
- Present GOTCHAS and PATTERNS
- Load relevant DECISIONS
Reflect Mode
- Extract patterns from completed work
- Identify gotchas from review failures
- Record architectural decisions with rationale
- Persist to .beads/knowledge/
Tool Use
Invoke via babysitter process: methodologies/metaswarm/metaswarm-knowledge-cycle
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- yesterday First seen · 45 lines · 33 tokens per session scan A 574d73247254
knowledge-curation is a skill published in the GitHub repository a5c-ai/babysitter (1,765 stars, last pushed yesterday), licensed MIT. It adds 33 tokens to every session and 379 once invoked, about $0.0002 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 2026-09-03.
Other skills, from other repositories
swarm-init
Start a new project connected to the SwarmVault — register it, wire platform adapters, offer git init, set vault-as-default. Use when starting a new project, connecting a project to the vault, or when the user asks to set up SwarmVault in a directory.
swarm-migrate
Bring existing projects into the SwarmVault — register, mirror docs, optionally mine the repo into SDLC artifacts and resume the flow mid-phase (brownfield adoption). Use when the user wants existing projects migrated/connected to the vault, or an existing codebase placed into the SDLC flow.
swarm-eject
Disconnect a project from SwarmVault and choose what happens to its knowledge — keep it, export it into the repo, or delete it. Use when the user wants to stop using SwarmVault here, remove the vault wiring, uninstall the framework from a project, offboard or archive a project, or delete a project's vault data.
groom-backlog
Groom a project backlog in an issue tracker. Loads the whole board in bulk, computes a gap inventory instead of eyeballing one, verifies each candidate issue's factual claims against the code and the tracker, ranks the verified candidates by a stated four-tier heuristic, proposes an evidence-backed closure for an…
kano-backlog
Prioritize and refine a GitHub Issues backlog with the Kano model — categorize every open issue as Must-be, Performance, Attractive, Indifferent, or Reverse, apply Kano + priority labels back to GitHub automatically, and recommend the single best next issue to pick up. Use this whenever the user wants to triage…
declare
Analyzes task plan and declares files that will be created or modified, producing declared-files.json.