Repository navigation
Require answer evidence and block unsafe adversarial acceptances - #94
Draft
aryanorastar wants to merge 1 commit into
Draft
aryanorastar wants to merge 1 commit into
aryanorastar wants to merge 1 commit into
Conversation
This was referenced Oct 30, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Customer problem and change
Answer-affecting changes can omit comparison evidence, and the existing eight rule-only checks do not expose several unsafe model-review acceptances. Prepare a trusted-main metadata policy with an actual current-head check and add six original fictional acceptance probes to the no-model runner.
The policy covers production root code/data/dependencies, all scripts (including ingestion), prompts/filters and workflows. It requires a current-head comparison evidence link or a concrete no-answer-change justification with an exact-head, exact-section-digest, non-author maintainer approval. Review submission/edit/dismissal events use an unprivileged relay; the privileged main workflow freshly reads open PR metadata, never branch code or artifacts. Invalid/partial metadata and races block the check.
Verification and actual blockers
Installation and release
The workflow and required-check enforcement are not installed on main. Fresh reads show classic branch protection absent and effective main branch rules empty. After general acceptance repairs, accuracy disposition, review and green current-head checks, the trusted workflow must be merged and its fixed
Mira answer evidencehead check made required, then verified. No repository settings were changed.A comparison URL declaration is not machine grading of accuracy or evidence adequacy. Read and grade the linked record independently; the implementation documents its limits and does not claim a tamper-proof control against administrators or malicious workflow maintainers. docs/ANSWER_EVIDENCE_GATE.md has the staged installation checklist.
#90 is merged cffb46b but its provenance code does not enforce hardware applicability. Production stays the healthy a9c9508/a6ef6b22 rollback; release and the third Discord smoke remain held. No paid run, customer/protected data, Discord post, optional enablement, merge or deployment.
Answer-change evidence
Mode: pending
Head: eeaf75c
Justification: This is a blocking draft: two adversarial acceptance regressions remain unresolved and required-check installation has not occurred.