Kit Map Steward — Development agent for Claude Code
Use proactively when three or more map misses are unstewarded — the stop-gate names it (map M-21).
How to install Kit Map Steward
Installs to ~/.claude/agents/oskarc-meta-kit-builder-kit-map-steward.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/oskarc/meta-kit-builder/HEAD/agents/kit-map-steward.md -o ~/.claude/agents/oskarc-meta-kit-builder-kit-map-steward.md Restart Claude Code, or start a new session, for it to be picked up.
What Kit Map Steward does
name: kit-map-steward description: Use proactively when three or more map misses are unstewarded — the stop-gate names it (map M-21). Reads misses, firing telemetry and the map itself, desk-probes the map against real situations, and proposes entries to add, narrow, demote, split, retarget, retire or ratify. Writes proposals to the ledger only; the pioneer ratifies and the main agent edits MAP.md. model: sonnet effort: medium maxTurns: 25 tools: Read, Grep, Glob, Edit, Bash color: cyan hooks
Alternatives in Development
- Systematic Debugger — Use this agent when you need to diagnose and fix bugs, errors, or unexpected behavior in code 39.2k ★
- Store Steward — Reviews schema, migration and store-size consequences of a change 96 ★
- Po Steward — Steward seat on the Atlas PO Council 79 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Act 1 Setup
Read-only narrative steward for Act I (slides 1-5) of the Workflow SDK GA deck. Covers the cold open, happy-pa
Kit Case Clerk
Use proactively when pioneer corrections are unclerked — the stop-gate names it (map M-15). Turns recorded cor
Kit Consolidator
Use proactively when ledger observations are unconsolidated — the stop-gate names it (map M-14). The only writ
Dispatch Input Checker
Cold dispatch-input reality-check at the aif-dispatch station boundary. Given ONLY the dispatch input (kickoff
Act 2 Verbs
Read-only narrative steward for Act II (slides 6-14) of the Workflow SDK GA deck. Covers the three workflow sc
Shipped Agent Liveness Prober
Probes each shipped sub-agent for behavioural liveness via with/without-tools fresh-subagent dispatch, capturi
Related Skills
Dotnet Refactor
Phase-gated refactoring/design loop - session-blind subagents build the map and traces, empirical probes prece
AegisOps
Agentic AI copilot for service observability: synthetic multi-service telemetry, three competing anomaly detec
5 Whys Dtd
Ask why until the cause is actionable: three to five links, each answering the last, and an intervention that