llm-deepseek gains a Loader+Include composition spec proving external
settings.yaml/.env edits reach the very next request, and a real-API e2e
where only a credentials-local document holds the key. The headless example
pins the first-run missing-credential UX as a keyless stream-json snapshot
(new credentials.cordis.snapshot.yml scenario); runLoaderSmoke learns
expectedExitCode so a designed failure surface can be pinned instead of
masked.
The TUI keyless smoke exposed the wrong idle signal: /plan submitted while
the driver is between turns still saw status running (checkpointing), so
the selection queued and the immediate copy never appeared. The commit
gate is now the log's own open-turn fold — plan/mode commits at once
whenever no turn is open, and queues only inside one. Specs drive the
distinction through explicit turn/start / turn/end events instead of a
fake status; the keyless smoke pins the new idle copy on both /plan and
/plan off; regenerated cordis/service catalogs ride along.
The idle /plan and /plan off command paths pin the immediate-commit copy;
the chip's dead leaving/locked guard is deleted (both states disable the
button, so no click reaches the handler); the multi-turn TUI golden is
re-recorded keyless over the idle-commit wording and the standalone
plan/mode narration ordering.
The ACP fixture configured read-only as the deployment default for both parent and child. Its delegated write therefore remained denied even with inheritance disabled, so the scenario could not fail on the regression it claimed to protect.
Delete the overlay, scenario registration, sessions, prompt, and 473-line tool-schema sidecar. The Loader-booted headless snapshot remains the real composition guard: only its parent carries read-only while the deployment default is workspace-write, so removing inheritance makes the child write reach disk and fails the test.
Keeping one discriminating snapshot avoids 662 lines of duplicated fixture data and makes the review evidence correspond to the actual security boundary.
The parent implementation introduced sandboxMode and approvalPolicy as generic SessionHeader fields, then propagated those fields through both persistence backends, session-query indexes, collision checks, policy-specific seed-boundary folds, catalogs, and a broad test matrix. That storage plane is unnecessary: Session already accepts a validated constructor seed, and persistence captures that seed when the session is announced before committing its first batch.
Capture each parent override synchronously at delegation, append source-tagged sandbox/mode and approval/policy records after the optional fork prefix, and create the child with that combined seed. Keeping header.seedLength at the original fork-prefix length preserves lineage while ordinary last-event-wins folds make the inherited records outrank stale parent history and remain subordinate to later child switches. Unswitched parents still stamp nothing, so children continue to follow deployment defaults.
Remove the generic header fields and every persistence/query/schema branch built around them. Collapse the inheritance suite from ten leaking scenarios to four owned-context cases covering real filesystem confinement, stale fork precedence, delegation-time capture, and the no-override path. The assembled headless snapshot now asserts the persisted inheritance event directly.
This keeps the security behavior while restoring policy ownership to the existing event log and deleting the speculative durability machinery that the original tests did not exercise.
Retarget the feature branch to the current master tip without rewriting its existing review history. Keeping this as a dedicated merge checkpoint makes the later simplification diff attributable to the stacked child rather than mixing base movement with design changes.
Resolve the identified-message API drift in the feature tests by constructing complete user messages, reading the nested tool-result message shape, and adapting the prompt-submit listener signature. Preserve both sides of the user-approval conflict: master’s createUserMessage wrapper and the feature’s inherited-policy attribution.
Regenerate the Cordis and persistence catalogs, re-record the session README pair, and refresh the affected ACP/headless fixtures so derived artifacts describe the merged source rather than either parent in isolation.
Validated with the focused policy/session/persistence/query suites (430 tests), focused ACP/headless snapshots (3 tests), build, doc-sync (25 gates), lint, hygiene, and git diff checks.
Boot the glob sampling scenario from a standalone ACP composition that exposes only bash, glob, and grep. Regenerate the smaller header fixtures and trim implementation narration already owned by the Agent Note.
Remove the duplicate model-facing list tool from this branch; directory orientation remains available through bash ls. Keep the glob sampling fix, add a real ACP composition snapshot, and narrow the decision record to the shipped bug fix.
Yichen Jiang
ZiyaZhang
Turtle
Tianyi Cui
imccyu
_Kerman
kingwl
NI0317