fix: provision dependency-only directories safely

[ruby-dlee]

Intent

Fix the shared worktree-provisioning classifier in bin/fm-provision-lib.sh so a directory carrying only a dependency list installs those dependencies and is never treated as a runnable package. Establish and fix the actual failing command rather than assuming the reported requirements.txt diagnosis; preserve an honest capability-gap launch when safe dependency-only installation is unsupported. Cover the direct execution path with regression tests that prove requirements-only directories never become package installs and inspect sibling lockfile-only manager cases. Confirm that pyproject.toml without a dependency lock/list, ambiguous managers, below-scan-depth, and over-budget outcomes are deliberate separate policies. Do not silently change the fleet-wide component budget; include a reasoned, evidence-based recommendation for relvino's observed 23 components in the completion report, with any proposed default change left as a separate change. Validate review, tests, documentation, lint, push, PR, and CI to green.

What Changed

• Treat bare requirements.txt and resolvable standalone uv.lock files as dependency lists without installing their directories as packages; report unsupported Python and JavaScript lockfile-only cases as capability gaps.
• Harden Herdr event-reader environment isolation and distinguish unrelated fleet churn from lab-attributable tripwire failures.
• Document the provisioning policies and expand regression coverage, including isolated test homes and configuration fixtures.

Risk Assessment

✅ Low: The change is well-bounded, preserves requirements.txt precedence and fleet-wide limits, and conservatively converts unsupported standalone lock shapes into explicit capability gaps.

Testing

The reported baseline and focused end-to-end CLI tests all passed, demonstrating dependency-only provisioning, separate pyproject/manager/depth/budget gap policies, guarded Herdr drift classification, and live Herdr spawn/event behavior; reviewer-visible CLI transcripts were captured and the worktree remained clean.

Evidence: Provisioning end-to-end transcript

== tests/fm-spawn-provision.test.sh ==
ok - a worktree declaring no recognized manifest provisions nothing and succeeds
ok - a bare requirements.txt installs only its dependencies
ok - bare requirements take precedence over a standalone uv.lock
ok - a standalone uv.lock installs only its locked dependencies
ok - an unresolved standalone uv.lock records a capability gap without refusing
ok - a path-dependent standalone uv.lock records a capability gap
ok - an npm lock without its required manifest records a capability gap
ok - a pnpm lock without its required manifest records a capability gap
ok - provisioning installs once and reuses an unchanged worktree without install cost
ok - a changed dependency manifest invalidates the cache
ok - installer configuration that changes what is installed invalidates the cache
ok - a fingerprint match with a broken environment is a miss, not a hit
ok - Node installs include declared validation dependencies deterministically
ok - an environment whose recorded runtime no longer matches is reprovisioned
ok - pip directives and includes are parsed, so an unchanged component stays cached
ok - build output written inside node_modules does not invalidate a healthy cache
ok - a declared manifest that cannot be read refuses instead of hashing to nothing
ok - a component nested past the scan depth is reported everywhere, not dropped
ok - a dependency scan that outlives its bound warns, reports, and launches unprovisioned
ok - a uv pip check that could not run is recorded as unverified, never as a verdict
ok - a python project with no committed lock or requirements file is named, not silently skipped
ok - the pyproject declaration read is bounded and degrades to undetermined
ok - a pyproject.toml holding only tool configuration provisions nothing and reports nothing
ok - a reported metadata inconsistency is recorded and launches; an unusable environment still refuses
ok - a failing install refuses the spawn and records no fingerprint
ok - an install that hangs is bounded and refuses instead of wedging the spawn
ok - a signal-killed bounded install is reported as a failure
ok - empty, zero, and nonnumeric provisioning tunables refuse before detection
ok - a traversal failure is distinct from a successful undeclared-worktree no-op
ok - a directory name containing another as a token is not deduped away
ok - a bounded call to a command that cannot be executed is a failure
ok - an environment signature that captures nothing refuses instead of hashing nothing
ok - a host without python3 warns and launches unprovisioned rather than refusing
ok - a host with no bounding mechanism runs nothing and launches unprovisioned
ok - a python project with no uv is left unprovisioned and reported, not refused
ok - a recognized but unsupported package manager is a recorded gap, not a refusal
ok - a worktree over the component budget provisions what it can and reports what it skipped
ok - the component budget is spent on what the task names before detection order
ok - a host capability gap is reported to the lane, not only to the operator
ok - a symlink planted at the report path is replaced rather than written through
ok - a requirements graph past the traversal cap is a gap, never a fingerprint over a prefix
ok - the JS package manager comes from what the project declares, not from lockfile precedence
ok - detection is driven by what the worktree declares and skips installed trees
ok - a uv workspace syncs every member while a single project keeps the plain sync
ok - uv workspace member manifests participate in the cache fingerprint
ok - a Node pin is read from what the project declares, and only when unambiguous
ok - a declared Node runtime that cannot be found leaves the JS components unprovisioned
ok - a declared Node runtime is resolved to its highest installed match and exported for the lane
ok - a Node pin wins for the Node toolchain and changes resolution for nothing else
ok - a shared pinned Node prefix is published once and never rebuilt under a running lane
ok - concurrent publishes each return a complete prefix and never nest one inside the other
ok - a published prefix that stops validating is stepped over, leaving the stale one untouched
ok - a spawn provisions its worktree after verification and before endpoint creation
ok - provisioning failure refuses the spawn instead of launching a lane that cannot validate
ok - a worktree over the component budget launches, warns, and records what it skipped
ok - the crewmate can read, inside its own worktree, what provisioning skipped and why
ok - teardown removes the provisioning log with the rest of the per-task state
ok - a later provisioning failure leaves the worktree clean without writing to the main clone
ok - an unignored install directory refuses before any installer runs, and never writes to the main clone
ok - provisioning is opt-out per spawn and per home, and the choice is recorded
ok - a lease that provisions nothing clears the previous task's report instead of inheriting it
ok - an unparseable provisioning setting refuses rather than silently disabling the gate
ok - a project declaring no recognized manifest spawns exactly as before
# all fm-spawn-provision tests passed

Evidence: Herdr tripwire behavior transcript

== tests/fm-herdr-lab.test.sh ==
ok - fm-herdr-lab: names fail closed and require the lab prefix
ok - fm-herdr-lab: provision timeout is configurable only within a bounded range
ok - fm-herdr-lab: provisioning, scoped calls, guarded teardown, and fleet tripwire are deterministic
ok - fm-herdr-lab: missing tripwire refuses teardown before any Herdr call
ok - fm-herdr-lab: agent argv routing inserts only the owned session before --
ok - fm-herdr-lab: changed default fleet state is a hard failure
ok - fm-herdr-lab: unrelated same-operator fleet churn is audited without blocking teardown
ok - fm-herdr-lab: lab-attributed default fleet drift retains teardown refusal
ok - fm-herdr-lab: transient default runtime activity does not change fleet identity
ok - fm-herdr-lab: provisioning requires a proven-running default session
ok - fm-herdr-lab: default running proof requires the JSON Boolean true
ok - fm-herdr-lab: destructive lab boundaries re-prove the running default session
ok - fm-herdr-lab: an owned stopped lab can re-provision safely
ok - fm-herdr-lab: failed deletion retains ownership until absence is confirmed
ok - fm-herdr-lab: timed-out provisioning cancels the launch before teardown

Evidence: Live Herdr backend smoke transcript

== tests/fm-backend-autodetect-smoke.test.sh ==
ok - real herdr: fm-spawn.sh auto-detects herdr from HERDR_ENV=1 (no explicit config) and prints the loud notice
ok - real herdr: auto-detected spawn records backend=herdr and herdr_session/workspace/tab/pane fields in meta
ok - real herdr: the auto-detected spawn's launch command actually ran in the herdr pane
ok - real herdr: teardown completes the auto-detected spawn/teardown cycle (meta cleared, pane closed)
== tests/fm-backend-herdr-eventwait-smoke.test.sh ==
ok - real herdr (herdr 0.7.3): events.subscribe capability gate passes (protocol >= 16, events surface present in api schema)
ok - real herdr (herdr 0.7.3): a driven idle->blocked transition returns the blocked record in 0.425s (pane w1:p2)
ok - real herdr: the watcher fast-path enqueues a stale wake naming the task window from the live blocked transition

- Outcome: 🔧 1 issue found → auto-fixed (7) ✅ across 8 runs (17h45m53s)

Pipeline

Updates from git push no-mistakes

✅ **intent** - passed

✅ No issues found.

✅ **Rebase** - passed

✅ No issues found.

🔧 **Review** - 1 issue found → auto-fixed (2) ✅

• 🚨 bin/fm-provision-lib.sh:1534 - The new standalone-uv.lock path creates a synthetic pyproject containing only the root name/version, then runs uv sync --frozen. A real lock generated from a project with dependencies will not match that synthetic metadata, so uv refuses to update under --frozen; the test fake only checks that pyproject.toml exists and therefore misses the real failure. This also leaves a directory containing both requirements.txt and a standalone uv.lock on the broken uv-lock path because detection prioritizes uv.lock. Preserve the required capability-gap launch for standalone locks unless matching project metadata can be reconstructed safely, and prefer the supported requirements installer when requirements.txt is present without pyproject.toml.

🔧 Fix: Prefer requirements over standalone uv locks
1 error still open:

• 🚨 bin/fm-provision-lib.sh:1535 - Standalone locks are classified as supported after validating only the root entry. A valid lock with root source = { virtual = "." } and a dependency sourced from a relative workspace/path remains reachable: copying the lock to a cache directory changes the relative-path base, while --no-install-workspace may omit that dependency entirely. The probe does not verify locked packages for uv-lock, so provisioning can either fail the whole spawn or report success with an incomplete environment instead of the required honest capability gap. At the classifier boundary, reject locks containing non-root relative/workspace sources unless their original resolution semantics can be preserved safely.

🔧 Fix: Reject path-dependent standalone uv locks
✅ Re-checked - no issues remain.

🔧 **Test** - 1 issue found → auto-fixed (7) ✅

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Isolate watcher guard configuration fixture
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Isolate watcher fixtures from live configuration
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Isolate test homes and reap Herdr event reader
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Isolate arm policy fixture home
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Confirm AFK failure was ambient Herdr interference
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Confirm bounded I/O cleanup passes focused stress
1 error still open:

• 🚨 tests failed with exit code 1
• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi

🔧 Fix: Audit external Herdr churn while preserving attributed tripwire
✅ Re-checked - no issues remain.

• if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi
• Outer-executor baseline: if [ "${FM_AZURE_VALIDATION_CELL:-0}" = 1 ]; then exec "$FM_AZURE_VALIDATION_SHARD_BRIDGE" behavior --count "${FM_AZURE_VALIDATION_SHARD_COUNT:-8}"; else exec bin/fm-no-mistakes-test-command.sh; fi (reported successful)
• bash tests/fm-spawn-provision.test.sh
• bash tests/fm-herdr-lab.test.sh
• bash tests/fm-backend-autodetect-smoke.test.sh
• bash tests/fm-backend-herdr-eventwait-smoke.test.sh
• Verified git status --short remained empty after testing

✅ **Document** - passed

✅ No issues found.

✅ **Lint** - passed

✅ No issues found.

✅ **Push** - passed

✅ No issues found.