fix(psychsocio): close post-merge P1 preregistration gaps

[erinepshovel-code]

Post-merge review of #29 found valid P1s despite a green path-scoped contract workflow. This repair stays inside the new stack-local psychsocio-metafauna research boundary plus the workflow that validates it.

Repairs:

• bind METAPAT to the stack-pinned research/metapat/BASE.json source (34d954aa...) instead of an unrepresented live commit, use the workspace-introduction stack identity to avoid nonexistent/self-recursive provenance, and recompute the work-graph digest;
• narrow H2 to a fully frozen held-out P1–P4 difference-in-differences interaction precursor; threshold/breakpoint coalescence remains UNRESOLVED regardless of H2;
• make H4 a two-cohort paired separation test: matched-support PM must preserve inspectability while paired high-demand Pi must actually satisfy frozen narrowing gates; empty eligible cohorts are UNRESOLVED and missing source episodes are BLOCKED;
• make H5 use exact held-out support/no-support pairs, require persistence + reproduction-allocation reductions and agency gains, and classify missing pairs as BLOCKED rather than FALSIFIED;
• run the psychsocio contract workflow when research/metapat/BASE.json changes and test that trigger;
• add regression checks for the exact authority, provenance, H2/H4/H5 decision rules, and workflow dependency.

Scope: four files only: the preregistration, work graph, contract tests, and .github/workflows/psychsocio-metafauna.yml. No simulator implementation, canon promotion, human classification, or imported repository mutation.

Validation required before merge: exact-head psychsocio-metafauna workflow success, completed Codex review on that same head, no unresolved P1/P2 review threads, unchanged base identity, coherent four-file scope, and mergeability.

hmmm: the synthetic run remains not-run; a threshold-shaped coalescence model still requires its own preregistration.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-05T12:57:46.260434Z	3ff3fed	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f5091a86b4

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[erinepshovel-code]

Maintenance audit found two additional merge blockers on current head f5091a86b4af1ad0c018a25cd0580e1fc6206411:

1. P1 provenance: WORK_GRAPH.json identifies the stack participant as The-Interdependency/stack@5b24db9...:research/psychsocio-metafauna/, but that path does not exist at 5b24db9... (contents lookup returns 404). Pin the stack-local research source to an exact commit where the path actually exists — normally merged research: add psychsocio-metafauna synthesis #29 6c3f94b1... as this repair's base — then recompute work_graph_sha256 and freeze that identity in regression coverage.

2. P2 preregistration determinism: H4 has no rule for an empty eligible cohort (PM episodes with adoption >= 0.75). Zero qualifying episodes must not pass/fail vacuously. Freeze an explicit outcome; UNRESOLVED is appropriate because the high-adoption separation prerequisite was not exercised, and add a regression assertion.

The path-scoped contract workflow is green, but these provenance/decision-rule gaps remain load-bearing. Do not merge until repaired and current-head review/checks settle.

[erinepshovel-code]

@codex review

Re-review current head after repairing the three P1s from the prior review. Please evaluate the frozen H2 comparable-cell estimator, the narrowed nonlinear-coalescence interpretation boundary, and the held-out paired H5 aggregation rule.

[erinepshovel-code]

Current-head check after the H2/H5 repair commits: 180565a3ea2239a6f07fac768d6649f476b60bb4 now has a green psychsocio contract run, and the three earlier Codex P1 threads are outdated by the repaired text. The two maintenance blockers above are still present on this head: the work graph still points research/psychsocio-metafauna/ at stack@5b24db9... where that path did not exist, and H4 still has no zero-eligible-cohort outcome. Keep #30 open until those two are repaired, the digest/tests are refreshed, and review is rerun on the resulting exact head.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 180565a3ea

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a9315bac24

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".