Extract the existing automatic Planner research methods into instruction-only research-auto, without changing the manual research package or owner tools, budgets, provenance and workflow authority.
Immutable v0.12.2 at 86cf1b6f8173754897ce759bf04ec991fa494911 includes source/domain selection, publication/version checking, inference/corroboration, already-admitted empirical evidence and query-refinement methods from RelayView's existing launch policy. Earlier tags remain untouched. RelayView pairs this release with versioned launch/recovery presentation that removes the core copies and preserves retained assignments.
Evidence and review
Before: new assertions for publication-date and corroboration methodology failed against the incomplete package. After: all 20 Python package tests pass on the exact final head, including the automatic-method boundary; reproducible package/archive checks and RelayView public-source verification pass for v0.12.2.
Final cumulative Standards and Spec reviews each report 0 actionable findings at 86cf1b6. Related RelayView PR: anodyine/relayview#1043.
Documentation disposition
updated — README describes the automatic/manual and owner boundary and the existing policy methods included in this package.
Ablation decision
not_required — equivalent existing methodology extraction; no retrieval/ranking, routing, budget or permission change.
Merge Danger
Door: two-way for future selection; public tags remain immutable.
Blast Radius: Planner instruction exposure only; no executable files or new tools.
Rollback: remove the package from future assignments while preserving public tags and captured assignments.
Expected: 20 passing tests and package identity 64dcdbc1d73f052d70ee01bf5362733d3a0ed006638744249195b8e20a0f3bd3. Stop on failure and retain output. These checks establish content/admission, not model compliance. Operator review of the extracted wording remains required.
Post-Merge Cleanup
Keep the worktree for review follow-ups until the linked RelayView change is accepted. Preserve all immutable release tags.
## Summary
Extract the existing automatic Planner research methods into instruction-only `research-auto`, without changing the manual `research` package or owner tools, budgets, provenance and workflow authority.
`core tool/policy methods → selected Planner research-auto`
Immutable v0.12.2 at `86cf1b6f8173754897ce759bf04ec991fa494911` includes source/domain selection, publication/version checking, inference/corroboration, already-admitted empirical evidence and query-refinement methods from RelayView's existing launch policy. Earlier tags remain untouched. RelayView pairs this release with versioned launch/recovery presentation that removes the core copies and preserves retained assignments.
## Evidence and review
Before: new assertions for publication-date and corroboration methodology failed against the incomplete package. After: all 20 Python package tests pass on the exact final head, including the automatic-method boundary; reproducible package/archive checks and RelayView public-source verification pass for v0.12.2.
Final cumulative Standards and Spec reviews each report 0 actionable findings at `86cf1b6`. Related RelayView PR: https://tea.kyle-merritt.com/anodyine/relayview/pulls/1043.
## Documentation disposition
updated — README describes the automatic/manual and owner boundary and the existing policy methods included in this package.
## Ablation decision
not_required — equivalent existing methodology extraction; no retrieval/ranking, routing, budget or permission change.
## Merge Danger
**Door:** two-way for future selection; public tags remain immutable.
**Blast Radius:** Planner instruction exposure only; no executable files or new tools.
**Rollback:** remove the package from future assignments while preserving public tags and captured assignments.
## Reviewer prerequisites and verification
Python 3; no model credentials or paid calls.
```zsh
python3 -m unittest discover -s tests -v
python -m tools.packages check packages/research-auto
```
Expected: 20 passing tests and package identity `64dcdbc1d73f052d70ee01bf5362733d3a0ed006638744249195b8e20a0f3bd3`. Stop on failure and retain output. These checks establish content/admission, not model compliance. Operator review of the extracted wording remains required.
## Post-Merge Cleanup
Keep the worktree for review follow-ups until the linked RelayView change is accepted. Preserve all immutable release tags.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Summary
Extract the existing automatic Planner research methods into instruction-only
research-auto, without changing the manualresearchpackage or owner tools, budgets, provenance and workflow authority.core tool/policy methods → selected Planner research-autoImmutable v0.12.2 at
86cf1b6f8173754897ce759bf04ec991fa494911includes source/domain selection, publication/version checking, inference/corroboration, already-admitted empirical evidence and query-refinement methods from RelayView's existing launch policy. Earlier tags remain untouched. RelayView pairs this release with versioned launch/recovery presentation that removes the core copies and preserves retained assignments.Evidence and review
Before: new assertions for publication-date and corroboration methodology failed against the incomplete package. After: all 20 Python package tests pass on the exact final head, including the automatic-method boundary; reproducible package/archive checks and RelayView public-source verification pass for v0.12.2.
Final cumulative Standards and Spec reviews each report 0 actionable findings at
86cf1b6. Related RelayView PR: anodyine/relayview#1043.Documentation disposition
updated — README describes the automatic/manual and owner boundary and the existing policy methods included in this package.
Ablation decision
not_required — equivalent existing methodology extraction; no retrieval/ranking, routing, budget or permission change.
Merge Danger
Door: two-way for future selection; public tags remain immutable.
Blast Radius: Planner instruction exposure only; no executable files or new tools.
Rollback: remove the package from future assignments while preserving public tags and captured assignments.
Reviewer prerequisites and verification
Python 3; no model credentials or paid calls.
Expected: 20 passing tests and package identity
64dcdbc1d73f052d70ee01bf5362733d3a0ed006638744249195b8e20a0f3bd3. Stop on failure and retain output. These checks establish content/admission, not model compliance. Operator review of the extracted wording remains required.Post-Merge Cleanup
Keep the worktree for review follow-ups until the linked RelayView change is accepted. Preserve all immutable release tags.