Harden the owner contract, evals, and release pipeline (3.0.1) - #96
Merged
Conversation
Delegation defers to the kernel's isolation rule instead of restating it, maintainer-only calibration advice leaves the runtime text, and the contract tests assert the boundaries rather than the wording. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Every case states per-arm activation expectations and links into the shipped contract; the evidence page records the corpus as UNVERIFIED. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…er capability Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Release 3.0.1. Full notes in
CHANGELOG.md.What changed, by group
plugins/skiphow/skills/skiphow/): authority has a provenance (owner and trusted host policy versus repository instruction files); a commit is made only where its path is known to be safe; a read-only review never becomes repair; completion is relative to the authorized destination; earlier runs' branches are inspected, not cleaned; delegation states an authority boundary, defers to the kernel's isolation rule instead of restating it, and routes model and effort by consequence and complexity;writing-for-agentsdrops unsupported generalizations. Covered bytests/test_contract.py.evals/,docs/evidence.md): twenty-four cases over eighteen fixtures and five arms, each linked into the shipped contract, with a semantic validator that rejects a case no arm could satisfy. Two maintainer-only cases removed..github/workflows/,SECURITY.md,scripts/check_hosts.py):validate(contents: read) runs checks and the pinned Codex validator and refuses a tag not matchingVERSIONor not contained inorigin/main;publish-release(contents: write) runs no repository code;deploy-sitedeploys the validated artifact. Manual Pages deploy requires an exact tag or SHA.SECURITY.mdcarries the per-capability host matrix. Covered bytests/test_workflows.pyandtests/test_hosts.py.docs/decisions.mdindexes live decisions; the deterministic validator stops failing on editorial shape and validates the hook on its safety properties;VERSIONand both manifests at 3.0.1.Checks run locally
python scripts/check.py: passpython scripts/check.py --pytest -q tests: 236 passedgit diff --check: cleanRepository settings changed for the release
build_typewas alreadyworkflow.github-pagesenvironment allowed only branchmain; av*tag deployment policy was added so the tag-triggereddeploy-sitejob can run.UNVERIFIED
validatejob of the release workflow.🤖 Generated with Claude Code