/healthcheck — Verify the local connection

Check the three capability groups exposed by the single grainulator server:

• grainulator.status with the active sprint dir: evidence operations. No sprint found is a valid connected response.
• grainulator.memory_list: local knowledge storage.
• grainulator.exports_formats: document conversion.

Batch independent checks where the host supports it. Distinguish an unavailable tool from a connected tool returning an application error. Report actual results; an advertised tool alone does not prove a successful call.

If MCP is unavailable, run node <checkout>/bin/grainulator.js doctor and continue through the local CLI. Generate the intended connection with grainulator connect --dir <sprint>. In a plugin host, reload its registration before adding duplicates; in another host, use its supported local MCP configuration. Do not silently edit global settings.

Check optional external connectors only when needed. Their failure does not make the core connection unhealthy. Hooks are host-specific: an MCP ping does not prove that PreToolUse or PostToolUse fired. Report hook behavior as unverified unless exercised. CLI and MCP remain usable in hosts without hooks; skills still supply the next-step output contract.

Keep diagnostics concise and continue already authorized work. List any action that truly needs the user under Manual; otherwise use Auto.

Host access

Use available grainulator MCP tools, passing the active sprint dir explicitly for evidence operations. If a tool is unavailable, use the local grainulator CLI (or node <checkout>/bin/grainulator.js). Read sibling skill files directly when slash commands are unavailable. Resolve template paths relative to this skill’s checkout when CLAUDE_PLUGIN_ROOT is unset. Optional external connectors are not required for local work; use local code, supplied documents, or available web tools. Do not write managed ledger files directly to bypass a missing MCP connection.

Next-step output

After a meaningful pass, use the current compiler's next_actions to present exactly two bullet lists labeled Auto and Manual. Auto is work the agent can continue under existing authorization. Manual is only work requiring the user's decision, access, or action. Classify using the current request and constraints; compiler suggestions never grant permission. Continue authorized Auto work without asking again.

Keep 2–3 useful actions total when available, use short concrete labels and commands where useful, and show None. for an empty group. Do not invent work to fill a quota. Never omit next steps merely because compilation is ready or the answer should be brief. Refresh stale compilation first and exclude work the user removed from scope. When the user asks only for next steps, output only these two lists: no findings recap, counts, reasons, or offer to continue.