Preserve complete source ancestry for PRs #3512, #3513, #3515, #3516, #3517 and #3518. Repair bounded recovery, alias compatibility and evaluator packaging gaps while keeping registry reuse inside the canonical loader. Verified actual hook, setup, dashboard, evaluator, sandbox and profile workflows plus full npm, Python, lint and supply-chain checks. ECC-043 preparation only; integration remains with the delivery owner.
compileContextProfile now validates its options, loads the registry and
calls compileContextPlan(registry, options), which compiles against a
registry the caller loaded in the same call. resolveTaskContext and
routingEntries load the registry once and compile from it, so the sources
are hashed once per call instead of twice.
On this repository (full@1, Windows 11, median of 5):
resolveTaskContext 1332 ms -> 645 ms, routingEntries 1321 ms -> 721 ms.
The registry never comes from outside the call: nothing caches it or
accepts one from a caller, so a stale or forged registry cannot be passed
in. The two "Registry changed during ..." checks compared the two loads and
are gone with the second load; each call now works from one snapshot, and
skill bytes are still verified against it when read. compileContextProfile
keeps its signature, output and error order.
Split out of #2945, whose prompt hook is superseded by #3453.
resolveTaskContext and routingEntries each compile a plan, which loads and
hashes the whole registry, then load and hash it again for their own use.
Opening one skill's policy file counts the loads: three opens per call on
main (two registry hashes plus the admission read), where one load gives two.
- New: resolution and the routing index hash the registry sources once per
call (resolveTaskContext and routingEntries each open the policy twice).
- The policy-drift test injects its change on the open after the registry
is hashed. With one load that is the second open, not the third; its
assertion that the drift is rejected is unchanged.
Against main: 3 failures in tests/lib/context-selection.test.js.
- csv-export-delimiter's grader adds a fourth check: the data rows are
the exported values, with an optional header (the hidden intent says a
header is fine either way). Reference scores 1; both decoys 0.75.
- graderIsSound grades every decoy-* overlay and is sound only when each
scores below 1; --check-graders and the pre-billing check print the
decoy scores through formatGraderScores.
- isValidTrial and completedTrialKeys move to lib.js; coverage.js and
run-intent.js use them from there.
- run.js resumes with completedTrialKeys, so a provider-error or timed-out
key is retried, and writes one transcript per attempt
(<key>__<attemptId>.jsonl, recorded on the row) so a retry never
overwrites the failed attempt's record.
- armSummary computes trials, passes and means over valid rows and counts
the failed attempts apart as providerErrors.
Docs: question-effectiveness.md states the decoy rule and that a
csv-export-delimiter pass in the reported runs reflects the dialect only;
its Measures section and the harness README describe retries, per-attempt
transcripts and the hookless-trial stop.
Three #3481 review findings still hold on main 8ce348fb:
- csv-export-delimiter's grader checks only the dialect (CRLF, semicolons,
quoting), so a constant '"x";"y"' export or one emitting the column
names scores 1. Add decoy-constant and decoy-keys overlays and pin that
each scores below 1 and that graderIsSound reports them; a scenario whose
decoy passes is not sound.
- The pilot runner (run.js) marks errored keys done, so they are never
retried, and armSummary counts provider-error and timed-out rows in
trials, passes and means. Pin the summary on valid rows only, and drive
run.js end to end: a provider error is retried on resume, keeps its own
row and transcript, and the summary reports it apart.
- Disabling run-intent.js's abort on an unobserved hook left every test
passing. Drive both runners end to end with a gated trial whose hook
never runs: each must stop and record nothing for it.
fake-claude.js gains FAKE_CLAUDE_SKIP_HOOK and FAKE_CLAUDE_PROVIDER_ERROR.
The intent end-to-end setup moves into syntheticIntentRunner, shared with
the existing resume test.
Against main: 2 intent failures (the decoy cases) and 2 pilot failures
(summary, retry). The two abort tests pass on main, since both aborts
exist; each fails when its runner's abort is disabled
(if (false && !row.hookObserved)).
Issue #3492: scope migration aborted with generic verification errors in
two real states. When the destination scope is installed but inventory
reports enabled:false because the user explicitly set
enabledPlugins["ecc@ecc"]: false, setup now throws a targeted
DESTINATION_EXPLICITLY_DISABLED error naming the setting and file and
offering the native claude plugin enable recovery, leaving the user's
choice untouched. The generic not-enabled error also gains the enable
command in its recovery list.
uninstallSource now accepts a destination-only state when the source
scope is gone from the fresh inventory even if the uninstall command
errored, and SOURCE_UNINSTALL_FAILED recovery no longer suggests
uninstalling a source scope the inventory no longer lists; it prepends
claude plugin list --json so the stuck state can be inspected.
Tests: node tests/lib/claude-scope-migration.test.js — five new cases
cover explicitly disabled resume and single-scope destinations across
user/project/local settings files, the generic fallback keeping the
enable path, an uninstall error with the source already absent, and
non-circular recovery. 16 -> 21 passing; red against unpatched scripts
is 15 passed / 6 failed. Full-suite failure set is byte-identical to
unpatched HEAD in a zero-dependency environment.
Also sync the pinned SOURCE_UNINSTALL_FAILED recovery in
tests/scripts/setup.test.js ('migration JSON failures retain phase,
scopes, and exact recovery') with the new claude plugin list --json
diagnostic: 29 passed / 1 failed on the patched tree before the sync,
30 passed / 0 failed after, identical to clean 2d515e4f.
CodeRabbit security finding on the merged sed work (CWE-693, verified live
on main via #3495): GNU sed labels for b, t, T and : end at ';' or newline —
verified against GNU sed 4.9 ('bar;e cmd' parses label 'a' and still runs
cmd; '}' is not a label terminator). Reading those operands to end of line
let 'sed :a;e git commit --no-verify' pass the hook while sed executed the
e command.
Labels are now read up to ';' or newline and command parsing resumes after
them. Text, filename and quit operands (a, c, i, r, w, q, Q) keep the
end-of-line reading: nothing after them on the line executes.
5 regression tests pin the three bypass paths and the benign boundary
(a bare label and a bare branch stay data). Suite: 817/817.
haelyra