| name | check-init |
| description | Read-only health check of the memhub project in this repo — schema version, render freshness, write-log activity |
| framework | memhub |
| framework_version | 1.0.0 |
| last_updated | "2026-07-14T00:00:00.000Z" |
Verify this repo's memhub setup is healthy. Read-only — this skill
never writes, never fixes, and never runs memhub render itself.
This is the Codex counterpart to the Claude Code /check-init skill.
Both run the same read-only memhub commands; they differ only in
the agent identifier on whatever read-side telemetry the host
captures.
/check-init is now a thin wrapper: the health logic moved into
memhub doctor (issue #21) so the same checks — project, config, DB
integrity, retrieval/metrics, integrations — are available identically
from the CLI, a script, or any agent, not just this skill. This
skill's own job is just detection + reporting the doctor output.
Detection
Check 1 — .memhub/ exists.
- POSIX shell (bash/zsh):
test -d .memhub && echo "present" || echo "absent"
- Windows PowerShell:
if (Test-Path .memhub -PathType Container) { "present" } else { "absent" }
Run whichever matches the current shell — no bash or WSL required.
absent → report Red. Suggest /init-project if the user
wants to bootstrap memhub here. Stop here.
Check 2 — memhub binary on PATH.
- POSIX shell (bash/zsh):
command -v memhub >/dev/null 2>&1 && echo "present" || echo "absent"
- Windows PowerShell:
if (Get-Command memhub -ErrorAction SilentlyContinue) { "present" } else { "absent" }
Run whichever matches the current shell.
absent → report Red. .memhub/ exists locally but the CLI
isn't on PATH; tell the user to install or rebuild memhub. Stop
here.
Run the health check
memhub doctor --json
Parse the wrapped {"doctor": {...}} object:
overall — worst status across every check (ok / warn / error).exit_code — 0 unless an error-level check fired.counts — {ok, warn, error, skipped}.checks[] — each {id, group, status, message, detail?}, grouped
by group (project, config, integrity, retrieval_metrics,
integrations).
If memhub doctor --json itself exits non-zero for a reason other
than a reported error check (e.g. a crash), surface stderr and
report Red with that single finding.
If the installed binary doesn't support doctor yet (older build,
predates issue #21), fall back to memhub status and report on
schema version and row counts from that instead — the pre-doctor
behavior this skill used to implement directly.
Report
Summarize as one of:
- Green —
overall: ok. One-line summary: project name and the
counts line (e.g. "14 ok, 3 skipped, 0 warn, 0 error").
- Yellow —
overall: warn. List every warn check with its
message — most already name the fix (run memhub render, run memhub index rebuild, run /catch-up, ...). No writes are made
from this skill.
- Red —
overall: error, or .memhub/binary missing per the
Detection step above. List every error check and its message.
End the report with:
memhub --version output.- The
counts from doctor --json (or the human Summary: ...
footer from plain memhub doctor).
- A one-line "next action" recommendation if Yellow or Red.
Notes
- Read-only. Never write, never fix, never
memhub render from this
skill — memhub doctor itself never writes either (no table
writes, no fix-it mode; it only reports).
- This skill is user-level; it fires in any repo. In a repo without
.memhub/, it reports Red and points at /init-project.
memhub doctor --strict (not used by this skill) promotes warn
to a failing exit code for scripting; this skill already reads the
full checks[] list itself, so it doesn't need the exit code to
decide severity.- memhub's PRD principle "intentionally boring" applies here too:
this skill reports state. It does not heal, repair, or auto-fix.
