Molecule AI Dev Engineer B (MiniMax)
06edd764e2
Harness Replays on PR #2935 (head 08e9033e) failed in step
"Pre-clone manifest deps":
fatal: Remote branch 950d39a490c12ba0f355ed8ca03b23fda9884823
not found in upstream origin
Root cause: scripts/clone-manifest.sh's clone_one_with_retry()
branched on `$ref = main` and used `git clone --depth=1 -q
--branch "$ref"` for everything else. For SHA-pinned refs (the
whole point of RFC #2927 — pin every entry to an immutable
commit SHA), `--branch <sha>` fails: git's --branch only resolves
named refs, not SHAs. The pinned SHA exists in the repo
(verified via /api/v1/repos/.../commits/<sha>) but the clone
command never tries to fetch it.
Fix: add a 3rd branch — when `$ref` matches `^[0-9a-f]{40}$`,
clone the full repo (no --depth so the SHA is reachable in
history) then `git checkout <sha>`. Drop .git after checkout
to match the post-clone .git strip in clone_category().
Tested locally with MOLECULE_GITEA_TOKEN="" (anonymous clone):
30/30 repos cloned successfully, all 6 workspace_template
entries have config.yaml at their pinned SHAs (the load-bearing
completeness-precondition that PR #2935's
TestManifest_RefPinning_WorkspaceTemplatesIncludeConfigYAML
asserts).
CI impact: should turn the Harness Replays / Harness Replays
gate from RED to GREEN on PR #2935 — the pre-clone step is the
entry point for all downstream replays.
scripts/
Operational and one-off scripts for molecule-core. Most are self-documenting — see the header comments in each file.
RFC #2251 coordinator task-bound harnesses
There are three related scripts; pick the right one:
| Script | Purpose | Targets |
|---|---|---|
measure-coordinator-task-bounds.sh |
Canonical v1 harness for the RFC #2251 / Issue 4 reproduction. Provisions a PM coordinator + Researcher child via claude-code-default + claude-code templates, sends a synthesis-heavy A2A kickoff, observes elapsed time + activity trace. |
OSS-shape platform — localhost or any /workspaces-shaped endpoint. Has tenant/admin-token guards for non-localhost runs. |
measure-coordinator-task-bounds-runner.sh |
Generalised runner for the same measurement contract but with arbitrary template + secret + model combinations (Hermes/MiniMax, etc.). Useful for cross-runtime variants without modifying the canonical harness. | Same as above (local or SaaS via MODE=saas). |
measure-coordinator-task-bounds.sh (in molecule-controlplane) |
Production-shape variant that bootstraps a real staging tenant via POST /cp/admin/orgs, then runs the same measurement against <slug>.staging.moleculesai.app. |
Staging controlplane only — refuses to run against production. |
See reference_harness_pair_pattern (auto-memory) for when to use which
and the cross-repo design rationale.
Common safety pattern across all three
- Cleanup trap on EXIT/INT/TERM auto-deletes provisioned resources.
DRY_RUN=1prints plan + auth fingerprint, exits before any state mutation. Run this before pointing at staging or any shared infrastructure.- Non-target guard refuses arbitrary endpoints (the controlplane
variant is locked to
staging-api.moleculesai.app; the OSS variant requires explicit auth + tenant scoping for non-localhost PLATFORM). - Cleanup failures emit
cleanup_*_failedevents with remediation hints; no silenced curl. ADMIN_TOKEN expiring mid-run surfaces as a structured event rather than a silent leak.
Activity trace caveat
If activity_trace.raw == "<endpoint_unavailable>", the per-workspace
/activity endpoint isn't wired on the target build — the bound
measurement is INCONCLUSIVE on the platform-ceiling question. Either
wire the endpoint or replace with the equivalent Datadog query. Note
that /activity accepts a since_secs query parameter; see the
endpoint handler for the supported range.
Other scripts
cleanup-rogue-workspaces.sh— emergency teardown for leaked workspaces. Prompts for confirmation. Pair with the harnesses if a cleanup trap fails (seecleanup_*_failedevents).staging-smoke.sh— quick smoke test for the staging canary fleet (formerlycanary-smoke.sh).dev-start.sh— local-dev platform bring-up.
The rest are self-documenting in their header comments.