JoelClaw Capabilities Map (AF30-002)

Canonical discoverability surface for agent workflows.

Source of truth command:

joelclaw capabilities

Flows

Flow ID	Category	Goal
`system-health`	operations	Determine whether joelclaw is healthy now
`operator-signals`	operations	Lease credentials, write structured logs, and notify gateway
`run-failure-triage`	diagnostics	Diagnose failed runs and identify next repair step
`deterministic-recovery`	diagnostics	Apply dry-run-first deterministic runbooks for known error codes
`event-delivery`	operations	Send an event and verify function-run delivery
`gateway-operations`	gateway	Operate the always-on gateway session
`memory-health`	memory	Verify memory pipeline quality and retrieval behavior
`otel-observability`	diagnostics	Diagnose incidents from OTEL without pod logs
`loop-lifecycle`	automation	Start/monitor/manage autonomous coding loops
`search-and-vault`	memory	Locate canonical context across search + vault

Output is JSON envelope (ok, command, result, next_actions).
next_actions commands use template syntax (<required>, [--flag <value>]).
params entries provide pre-filled values/defaults/enums.
Catalog includes capabilityContract metadata: configured capability adapters, config paths, and registry entries.
Capability config precedence: flags > env > project .joelclaw/config.toml > user ~/.joelclaw/config.toml > defaults.
This map is documentation; runtime discoverability must come from joelclaw capabilities.