fix: MiniCroft boots a lean default pipeline instead of every installed matcher

[JarbasAl]

🤖 Auto-generated by Claude Sonnet 5 via Claude Code — NOT human-reviewed. Verify before acting.

CI installs ovos-core[plugins] so Adapt/Padatious are AVAILABLE (correct — see gh-automations#99), but that surfaced a separate problem in MiniCroft: IntentService.handle_reload_pipelines instantiates every installed opm.pipeline plugin regardless of intents.pipeline — that config only orders/selects among matchers already loaded, it never gates which plugins get constructed. With roughly fourteen pipeline plugins installed (m2v, m2v-prototype, persona, OCP x2, common_query, ...), MiniCroft ended up eagerly booting all of them every time, even though it only ever set intents.pipeline to a much shorter list. One of those plugins, ovos-m2v-pipeline, does a synchronous sleep(3) inside its intent-sync handler; combined with several heavy skill test suites in the fleet, this routinely pushed MiniCroft's 60s READY cap and made setUpClass error out.

The fix is intents.blacklisted_pipelines, not intents.pipeline: that is the config IntentService actually checks before constructing a plugin. MiniCroft now defaults to a new LEAN_DEFAULT_PIPELINE — Stop, Converse, Adapt, Padatious, Padacioso and Fallback, high and medium confidence tiers — which is the matcher set the ovoscope test suite itself ever asserts against (grepping test/unittests/*.py turns up no reference to the -low tier or to common_query; Stop is kept unconditionally because skills assert stop behavior fleet-wide). Before IntentService is constructed, MiniCroft now computes and sets intents.blacklisted_pipelines to every installed pipeline plugin the chosen pipeline doesn't cover, and restores it in stop().

Two escape hatches keep this additive rather than restrictive. A heavyweight suite that needs one extra matcher (say, m2v) can pass extra_pipelines=M2V_PIPELINE instead of restating the whole lean list, and default_pipeline= remains a full override when a test wants total control (e.g. Adapt-only, or the wider DEFAULT_TEST_PIPELINE). To make sure a lean config can never quietly mask a real problem, get_minicroft() now raises RuntimeError naming any pipeline id that failed to load — whether it's simply not installed or errored during construction — instead of proceeding with fewer matchers than the test expects.

Fail-before: reverting only ovoscope/__init__.py while keeping the new tests produces a collection ImportError (LEAN_DEFAULT_PIPELINE is undefined) for the new TestMiniCroftLeanBootDefault class in test_minicroft.py and for the new test_lean_boot_e2e.py; both pass once the fix is restored. A manual boot-time comparison in the same environment showed roughly 8.3s for the old "no blacklist, all plugins load" path against a real registered intent versus roughly 1.5–3.2s for the lean-default path with the same skill — the gap grows with every additional heavy plugin installed, which is exactly what pushed the fleet past the 60s cap.

The full test/unittests suite passes 548, with 9 pre-existing failures from a missing optional ovos-dinkum-listener dependency and 119 skips — identical to the failure set on dev before this change, confirmed by running the same subset against dev directly. docs/minicroft.md and docs/usage-guide.md document the lean default and both override mechanisms.

Base note: dev at the time of this branch includes #154 (skills-trained quiet window wait), which this change leaves untouched. PR #156 (bumping the ovos-core floor for padatious's move from [lgpl] to [plugins]) is still open, not merged, so this branch targets the current [lgpl,plugins] floor.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

At your service! I've gathered all the check results for you. 🫡

I've aggregated the results of the automated checks for this PR below.

🏷️ Release Preview

I've checked the 'Platform Support' matrix. 💻

Current: 1.6.16a1 → Next: 1.6.17a1

Signal	Value
Label	(none)
PR title	fix: MiniCroft boots a lean default pipeline instead of every installed matcher
Bump	build

✅ PR title follows conventional commit format.

---

🚀 Release Channel Compatibility

Predicted next version: 1.6.17a1

Channel	Status	Note	Current Constraint
Stable	⚪	Not in channel	-
Testing	❌	Too new (must be <1.0.0)	ovoscope>=0.7.2,<1.0.0
Alpha	✅	Compatible	ovoscope>=1.6.15a1

🔒 Security (pip-audit)

Looking for any weak links in the supply chain. ⛓️

✅ No known vulnerabilities found (121 packages scanned).

🔍 Lint

Checking if we've met all our check criteria. ✅

❌ ruff: issues found — see job log

⚖️ License Check

Legal eagle here! Checking those licenses. ⚖️

✅ No license violations found.

Policy: Apache 2.0 (universal donor). StrongCopyleft / NetworkCopyleft / WeakCopyleft / Other / Error categories fail. MPL allowed.

📊 Coverage

The coverage detectives have finished their sweep. 🕵️‍♀️

⚠️ 60.0% total coverage

Files below 80% coverage (15 files)

File	Coverage	Missing lines
ovoscope/simple_listener.py	0.0%	63
ovoscope/tts_intelligibility.py	0.0%	200
ovoscope/version.py	0.0%	5
ovoscope/classic_listener.py	18.2%	117
ovoscope/intent_cases.py	22.3%	153
ovoscope/pytest_plugin.py	47.0%	210
ovoscope/cli.py	47.8%	132
ovoscope/ocp.py	47.9%	61
ovoscope/e2e.py	53.5%	66
ovoscope/media.py	56.7%	101
ovoscope/listener.py	57.1%	127
ovoscope/voice_loop.py	59.0%	119
ovoscope/audio.py	63.4%	126
ovoscope/__init__.py	65.8%	371
ovoscope/media_provider.py	67.6%	23

Full report: download the coverage-report artifact.

🔨 Build Tests

Everything is bolted down and ready to go. 🔩

✅ All versions pass

Python	Build	Install	Tests
3.10	✅	✅	✅
3.11	✅	✅	✅
3.12	✅	✅	✅
3.13	✅	✅	✅
3.14	✅	✅	✅

---

Processing completed in 0.0001 bot-seconds ⚡