Make command history user-facing while preserving internal trace

[codeforester]

Goal

Make basectl history read as a history of user-invoked Base commands while preserving the internal execution trace for troubleshooting.

Background

The current history writer records every Python implementation process launched by a Bash command. A single user action such as basectl setup banyanlabs can therefore produce setup, bootstrap, route, and projects rows. basectl test and basectl build can show internal trust or projects rows without a corresponding test or build row. The result is technically accurate but confusing as a UX surface. The project-attribution fixes in #1668, #1669, and #1670 make the remaining event-boundary problem more visible.

Scope

• Define a history event boundary for primary user commands versus internal delegated steps.
• Record one primary history entry for each public basectl command that executes work.
• Mark delegated resolver, routing, bootstrap, and trust-gate records as internal child events with parent linkage where practical.
• Hide internal events from default basectl history output while retaining them in the local history data and logs.
• Add an explicit trace/include-internal view for diagnostics and JSON consumers.
• Cover setup, check, doctor, test, trust, and build flows, including failed trust/build paths.
• Update command help and documentation for the new history behavior.

Acceptance Criteria

• Default history for a setup/check/doctor/test/trust/build sequence shows the user commands rather than helper commands such as projects, route, and bootstrap.
• Directly invoked public commands such as basectl projects and basectl trust allow remain visible as primary events.
• Internal events remain available through an explicit trace/include-internal option and in JSON output when requested.
• A failed internal step is attributed to its primary parent command without losing the child log path or exit status.
• Existing project, manifest, status, exit-code, redaction, and log-path fields remain correct.
• Existing history records remain readable; the change is forward-compatible with older records that lack event-scope metadata.

Validation

• Focused Python tests for base_cli.history, base_history, and affected Bash command suites.
• Regression tests for setup, check, doctor, test, trust, and build history output.
• git diff --check, focused pylint, and the full ./bin/base-test suite.
• Hosted CI checks for the resulting pull request.

Non-Goals

• Do not delete or rewrite existing history records.
• Do not remove per-step log files or reduce diagnostic detail in logs.
• Do not change command execution, trust policy, project resolution, or setup semantics.
• Do not add a separate persistent tracing service or remote telemetry.

Project Fields

• Priority: P2
• Size: M
• Area: CLI
• Initiative: Contract Hardening
• Milestone: v1.8.0

Agent Assignment

Human first; implementation is approved for the current Codex session after issue and Project intake are complete.