Repository navigation
Description
Problem
The suites at tests/skills/rebase/scripts/rebase-workflow.bats:7 and tests/skills/rebase/scripts/marker-scan.bats:3 contain 13 cases covering replay evidence and marker dispatch. The scanner already has native and Unix shim tests, but that does not prove every legacy fallback and failure-through-caller case is redundant. Retirement needs a case-by-case ledger after the Go workflow lands.
Acceptance criteria
- AC1: Given every case in the two suites, when the retirement ledger is read, then each is mapped to a covering Go test or a named reason to retain it.
- AC2: Given a deleted case, when its ledger entry is inspected, then it identifies equivalent Go coverage for every still-shipped behavior that case protected.
- AC3: Given the retirement diff, when it is inspected, then only ledger-approved duplicate cases and their unused test scaffolding are removed.
- AC4: Given a covering test, when its ledger entry is read, then its native or shell-integration coverage kind and executing platforms are stated.
- AC5: Given each promised platform, when the covering tests run in CI, then they execute without skips and pass.
- AC6: Given frozen fallback behavior without equivalent Go integration coverage, when retirement occurs, then its cases remain.
- AC7: Non-functional: retirement changes no executable implementation or compatibility rule.
Scope and non-goals
Scope: the two rebase Bats suites and their retirement ledger, after children 01 to 04 and 10.
Non-goal: adding missing production behavior, deleting fallbacks, rewriting the marker scanner or removing hook tests. Retained fallback cases can be implemented as Go-driven shell integration before retirement; if absent, keep the Bats cases. This child does not wait for N-2 merely to remove proven duplicate tests.
Approach
Start with all seven workflow cases and all six scanner cases. Audit internal/agentrepo/markers_test.go:61, internal/agentrepo/markers_internal_test.go:47 and internal/agentrepo/markers_shim_unix_test.go:38 rather than assuming identical names imply parity. In particular account for missing/old binaries, support probes, stderr-only startup errors and failure propagation through status, continue and verify. Label Unix shell integration honestly; it is not native Windows coverage. Legacy-algorithm cases still lacking Go coverage block children 13 and 14. Finish their separate test retirement here once coverage exists; do not bundle Bats deletion into those production-removal children. Any Bats case retained beyond removal must exercise a surviving shell entry rather than a deleted algorithm.
Dependencies
Blocked by: #895, #896, #897, #898, #904.
Existing scanner coverage is already on main.
Size
S. Forecast 300 code lines: at most 242 measured Bats code lines deleted plus 58 ledger-check/coverage adjustments. Any substantive new Go behavior belongs to a prior child. Prose allowance: 250 lines.
scc could not run in the isolated checkout. Measured baselines count non-blank lines other than full-line # comments, including Python docstrings and embedded code conservatively. Forecasts include implementation, tests, fixtures and registration; they are estimates, not measured future Go. Generated output is excluded only under packages/agents/claude/.apm/skills/, packages/agents/codex/.apm/skills/, .claude/skills/, .agents/skills/ and tokens.json. The source under packages/agents/source/ and copies anywhere else count. The final diff must be at most 800 code lines and 2,000 prose lines, with no exception; split before implementation if either forecast grows beyond its limit.
Verify plan
- AC1, AC2: enumerate all 13 case names and inspect the assertion-level mapping to existing Go tests.
- AC3, AC7: inspect the path and content diff; executable implementations and compatibility behavior must be unchanged.
- AC4, AC5: inspect coverage-kind/platform entries and CI logs proving execution of every promised test.
- AC6: run remaining Bats cases and Go-driven frozen-fallback cases on their declared shell-capable platforms; compare retained-case list with the ledger.
- Definition of done: D1 each AC is verified at the merged head, with the commands and results cited in the pull request's Verification. D2 an independent review approves the exact merged head, and every finding has a recorded disposition. D3 required checks pass with no gate loosened. D4 the leak scans the repository specifies are clean, and every post made for this work passes the per-post leak review before it is posted. D5 every commit is signed. D6 the documentation matches the change or the Verification says why there is no impact. D7 every commit has an
Assisted-bytrailer and the pull request has theprovenance:agent-assistedlabel. D8 the pull request stays within the size limits. D9 follow-ups are filed as issues and listed under Related. D10 the merge closes the issue through a closing reference, the board shows Done, and the finalStatus:comment names the merge commit.
Risks
Weaker Go assertions can silently lose fallback or failure coverage. Undo: revert the retirement commit to restore the exact Bats cases.
Operator authority
None.
Metadata
Metadata
Assignees
Labels
Type
Fields
Priority
Work type
Projects
- StatusShow more project fieldsBacklog
Activity
tbhb-agent commentedon Oct 27, 2026
Refinement: READY — Retire rebase Bats cases covered by Go tests
Model: gpt-6-astra
Child 11 Retire rebase Bats cases covered by Go tests
Verdict: READY
Checked: Prior assertion-level coverage and retained-fallback requirements remain consistent; unaffected by the corrections. Forecast remains 300 lines.
Gaps: None.
Split: None required.
Definition of done
The definition of done for this issue (Chore). Tick each item and give its evidence.
check-leaks.shwith the repository's private-strings list, and the leak review of each post.docs/, CHANGELOG or help text), or the Verification section says "No doc impact" and why.provenance:agent-assistedlabel.Status:comment names the merge commit.mise run checkis green, and so are the coverage floorsmise run cover-go-checkandmise run cover-py-check, whichcheckdoesn't run.packages/agents/source/commits whatmise run generate-skills, thenmise run skill-tokens, thenapm installrewrite, in that order, andcheck-generated-skills(part oflint) passes.docs/agent-skills-migration.mdor the adoption docs, or the PR's Verification says "No consumer impact" and why.Round: 2 of 3