adopt-project.py and shape-doctor.py: a plan answer that overrides a confident rule passes check and execute, then the doctor's first run says MISPLACED, and an emptied .gitmodules is asked on every run #655

Lane: openreposhape-1 (openRepoShape-1)

A plan answer that overrides a confident rule should not end adoption verified and then MISPLACED on the doctor's first run. And an emptied .gitmodules should have a record that can answer it, so the doctor stops asking about it on every run.

From lane 1's triage of 2026-10-15, D6, approved by Brett Heap at 2026-10-16T11:21:17Z ("Approve all (Recommended)"). One pull request covers the lines below.

Parent: #457 (#460, the doctor's tracking issue, is closed). Milestone: v1.1.0.

Lands after: #617.

Effort: M. Platform: all.

Lane openRepoShape-1 claims it at filing; see its CLAIMED comment.

Why this is P2.

• L1-57 is a wrong verified result. docs/ (rule spec-governance) answered code ends adoption verified, and the doctor then says MISPLACED (1 path), exit 1. The line measured it, the writer and the review, on main 2e6c6fa and on adopt-project.py, shape-doctor.py, validate-manifest.py: execute records the plan's answers under answered_paths: and a moved submodule names its row, so an answered examples/ is not asked again and a deliberate move names no plan (#525) #617's 3b78fae.
• L1-56. git rm of a leg's only submodule leaves a 0-byte .gitmodules. ambiguous-gitmodules asks about it on every run, and no record can answer it.
• The riders concern the same answered_paths: rows that adopt-project.py, shape-doctor.py, validate-manifest.py: execute records the plan's answers under answered_paths: and a moved submodule names its row, so an answered examples/ is not asked again and a deliberate move names no plan (#525) #617 adds. L1-58: a directory row without its / settles nothing. L1-59: a row also settles files added later, and nothing says so.

The open question for L1-57.

• (A) execute records an override of a confident rule as an answer the doctor honours, as adopt-project.py, shape-doctor.py, validate-manifest.py: execute records the plan's answers under answered_paths: and a moved submodule names its row, so an answered examples/ is not asked again and a deliberate move names no plan (#525) #617 records a question's answer.
• (B) check finds the override and asks for a resolution:.

Direction: A. The plan is the person's word, as #617 already treats an answered question.

Lines

• L1-56 (P2): an emptied .gitmodules (0 bytes, left by git rm of a leg's only submodule) is asked by ambiguous-gitmodules on every run, and no record can answer it: answered_paths: leaves it out and submodules: answers it only where the leg keeps one. Where: shape-doctor.py's ambiguous-gitmodules. The line measured it on 3b78fae and on main 2e6c6fa. Not re-checked since. (source)
• L1-57 (P2): a plan answer can override a confident rule: docs/ (rule spec-governance) answered code passes check and execute says adoption verified, but the doctor's first run says MISPLACED (1 path). The line measured it on 3b78fae and on main 2e6c6fa. Not re-checked since. (source)

Rides along (backlog; tick it here if the same change covers it)

• L1-58: a hand-written answered_paths: row for a directory without its trailing / (examples) passes the validator and silently settles nothing, because _covering reads it as a file. From the code at 3b78fae; the validator's silence was probed. Not re-checked since. (source)
• L1-59: an answered directory row also settles files added under it after the adoption, because source_commit is never consulted. By design, but nothing a row's reader sees says so. From the code at 3b78fae. Not re-checked since. (source)

Each line above is a line of #457's checklist comment (issuecomment-6200110061). Tick it there too when this closes.