test(periodic-report): align CLI cases with exact refresh contract

[mikamikasuki]

Goal And Delivered Outcome

• Outcome basis: The periodic-report protocol ties successor milestones to the exact refresh that accepted the successor Vision.
• Goal/source and gap: Three CLI tests still expected an ordinary later Todo completion to create or recover an intent after that boundary had passed.
• Observable before → after: On canonical main 647e216, the focused three-module set produced 7 failures and 6 passes. The updated tests pass on head 28a18f6 (13 passed) while preserving the valid exact-refresh path.
• Issue/task and intended base: No separate issue is required for this test-only contract correction. Target main at 647e216.

Author Declaration

• Written by: model_agent — OpenAI GPT-6 Luna Medium.

Implemented against

• Specification and revision: docs/reference/protocols/periodic-report-v0.md at 647e216, sections “Successor Vision milestones” and “Agent-semantic Goal Channel requests.”
• Criteria:

Criterion	Disposition	Test
A later Todo completion cannot report a superseded successor boundary	implemented	tests/cli_commands/test_periodic_report_intent_capability_chain.py; tests/cli_commands/test_todo_complete_settlement_capabilities.py
An accepted exact-successor refresh remains reportable with its capability context	implemented	tests/capabilities/test_periodic_report_native_successor.py

• Self-check: Reproduced the stale expectations on exact current main, checked the protocol and related issues/PRs, preserved the valid positive path, and reviewed the three-file test-only diff.

Scope And Continuation

• Completed scope: Corrected stale ordinary-Todo replay expectations and recovery counts while preserving exact-refresh coverage.
• Slice boundary: Complete within this scope; no production behavior changes.

Validation

• Tested revision: 28a18f6 (base 647e216).
• Run state: finished. Input classes: synthetic.

Check	Result	Evidence
Regression parity	passed	Exact current main: 7 failed, 6 passed; updated head: 13 passed across the same three modules.
Unit	passed	Focused periodic-report tests: 13 passed in 18.53s.
Architecture	passed	Project registry I/O census: 7 passed.
Static and repository checks	passed	Ruff, Python compilation, git diff check, DCO trailer, and semantic-vocabulary drift smoke.

• Coverage and gaps: The change updates tests only; no production runtime behavior is modified.

Frontend / Visual Evidence

• UI impact: none. Source data: synthetic.

Type of Change

• Test update

LoopX Area

• Capability or extension (providers, adapters, skills)

Technical Direction

• N/A; this test correction enforces the existing periodic-report protocol.

Shared-authority RFC fixture impact

• N/A; no authority schema or provider conformance changes.

Boundary Checklist

• No private state, credentials, raw traces, internal links, or local machine paths.
• Scoped to the periodic-report refresh contract.
• UI impact marked none.
• Every commit includes a DCO Signed-off-by trailer.

[mikamikasuki]

CI attribution for run 37838624566:

The JUnit artifacts from Python shards 1, 3, and 4 contain 10 failures. I ran those same 10 cases on this PR's base (f67a6d8) and head (9eee959) with Python 3.11.12; all 10 failed on both. I then ran them on canonical main (3ed5d6b) with Python 3.14.7 and saw the same failures. None of the failing test paths are among this PR's three changed files, so this run does not show a regression from this change.

The failing cases are covered by existing follow-ups: replan qualification in #5962, the Todo digest expectation in #5991, the actor read fixture in #5993, successor fixtures in #5994, and Chat cadence/exact-read fixtures in #5980.

At the latest status read, shards 1, 3, and 4 had failed, shard 2 was cancelled, and the workflow's checks job was still queued; the run is not terminal. I will refresh validation after the related baseline fixes land.