Antigravity Maintainer Batch Release
When to Use
Use this skill for repository-wide AAS maintenance, maintainer-side PR repair or merge batches, canonical synchronization, AAS Core or Workbench changes, protected releases, and hosted catalog or legacy redirect infrastructure. Do not use it for ordinary contribution work that does not require maintainer privileges or canonical convergence.
Protected-Main Contract
Treat the repository root containing this skill as pull-request-only:
- Read
AGENTS.md,.github/MAINTENANCE.md, and current maintainer docs before mutation. - Never commit or push directly to
main, even when the user says “push to main.” That phrase names the final target state. - Preserve unrelated dirty work. Use a clean temporary clone or a topic branch for maintainer changes.
- Use
npm run merge:batchfor accepted source PRs. Do not substitute a raw merge API, generic GitHub skill, or generic push helper. - Let
automation/canonical-repo-stateown generated artifacts and contributor-credit convergence after the source batch. - Use
release:prepareandrelease:publishfor releases. They never authorize a directmainpush.
Source Checks
Before changing anything:
- Fetch
origin/main; prove the clean maintainer checkout is onmainand equalsorigin/main. - Inspect live PRs, issues, discussions in scope, Actions failures, Dependabot, CodeQL, secret scanning, and
npm auditwhere relevant. - Confirm current scripts from
package.json; do not rely on remembered release behavior. - Capture user worktree status separately and keep those files out of maintainer commits.
Maintainer Sweep
-
Triage every open PR before editing.
- Separate valid source changes, repairable PRs, conflicts, generated-only noise, promotional links, and unsupported ownership/license changes.
- Review semantics, safety, provenance, risk labels, limitations, source credits, and changed-skill evidence.
- Prefer narrow maintainer repairs on the contributor branch when maintainer edits are enabled.
-
Validate changed skills truthfully.
- Run
npm run validate,npm run validate:references,npm run security:docs, changed-skill evidence, and the relevant tests. - Inspect semantics, safety, provenance, declared risk, limitations, and all tracked bundle files directly. Treat inferred risk labels and heuristic quality scores as non-authoritative; do not change a skill merely to satisfy a lexical signal.
- Inspect the
skill-reviewworkflow on the exact current head SHA. reviewmeans Tessl semantic review actually ran or a valid identical-content result was reused.manual-review-requiredmeans Tessl credentials or credits were unavailable, or Tessl did not produce a passing result. Perform the maintainer semantic review and attest with--reviewed-head <full-40-character-sha>.- Any non-passing Tessl outcome produces
manual-review-required; complete the semantic review and bind the judgment to the exact head instead of treating a heuristic score as merge authority. - Never report
manual-review-requiredas “Tessl passed.”
- Run
-
Run checks in parallel where independent.
- Use the repository validation, test, docs-security, source-credit, reference, warning-budget, and targeted app checks required by the changed files.
- Fix deterministic policy failures in the source; do not wait for them as if they were flaky CI.
-
Merge accepted source PRs in conflict-aware order.
-
Run a dry classification first when useful.
-
For changed skill content, review the exact head and run:
npm run merge:batch -- --prs <PR_LIST> --reviewed-head <FULL_HEAD_SHA> -
merge:batchmay normalize the PR body and close/reopen the PR. GitHub creates the replacement workflow runs asynchronously; the command must wait for and approve only post-reopen workflow/check-suite IDs. Older runs on the same SHA cannot satisfy or fail the fresh gate. -
The routine protected checks are
pr-policy,pr-evidence,source-validation, andartifact-preview. The retiredaas-v1-baselineworkflow is not a merge prerequisite and must not be awaited or approved during source or canonical-sync batches. -
If the PR head or base changes, discard stale evidence and rerun from a fresh
origin/main.
-
-
Converge canonical state once after the source batch.
- Wait for the protected
automation/canonical-repo-statePR. - Verify its managed-only diff, required checks, merge result, and the resulting
origin/main. - If an unmanaged repair remains, use a topic PR; never patch
maindirectly.
- Wait for the protected
Hosted Catalog and Legacy Redirect Bridge
Treat the current catalog and the legacy user-site bridge as one public system:
- Current catalog:
sickn33/agentic-awesome-skillsathttps://sickn33.github.io/agentic-awesome-skills/. - Legacy bridge:
sickn33/sickn33.github.ioathttps://sickn33.github.io/antigravity-awesome-skills/.
For SEO, indexing, Pages, redirect, or infrastructure changes:
- Change the generator and verifier in the source repository through a protected source PR and
npm run merge:batch. - Keep the legacy deployment managed allowlist exact:
.nojekyll,redirect-manifest.json, andantigravity-awesome-skills/**. Reject any unmanaged sync diff or PR file. - Preserve Google verification byte-for-byte and the Bing
msvalidate.01meta on the legacy root. Record both in manifest evidence. - Keep skill counts dynamic, but retain intentional curated sitemap locks. Version manifest contract changes and record source provenance.
- Let
legacy-redirect-sync.ymlgenerate or update the fixed automation PR. Bind a fresh verifier run to the exact target head SHA, validate its run identity and managed file set, publish the required status only after that proof, then use protected auto-merge. - Recheck source
mainbefore merge, request the legacy Pages build explicitly after bot-authored merges, and wait for the exact merged commit to be built. - Verify locally generated output byte-for-byte, then verify all live legacy/current redirect pairs for a full audit. Retry transient CDN failures with the full audit rather than accepting a partial probe.
- Prove idempotence with a no-drift sync: no replacement, PR, verification, or merge steps should run; Pages and live probes must still pass.
Keep both repositories on least-privilege Actions defaults (read) and require external actions to be pinned to full commit SHAs. When changing these settings or action versions, rerun source CI, CodeQL, Pages, and a legacy no-drift sync before declaring completion.
AAS Core Preview Acceptance
For AAS CLI, MCP, stack, catalog-cache, or Workbench changes:
- Use the current scripts declared in
package.json; do not resurrect retired evaluator, benchmark, tuning-gold, transaction-fault, race, or frozen-matrix gates as routine prerequisites. - Run the focused Core tests with
npm run test:aas-v1, the catalog integrity check withnpm run check:aas-v1-catalog, and the relevant Workbench tests/build when its contracts or copy change. - Keep MCP local, offline, read-only, bounded, and non-mutating. The coding agent inspects the project, searches and reads the complete catalog, and chooses the exact skill IDs. MCP searches, reads, validates agent-owned composition, and compares without scanning the repository or writing to it. Core must not rank, recommend, exclude, or disable skills; metadata is informational only.
- Keep
aas-stack.jsonfree of Core selection policy. It pins catalog identity, targets, goals, and the exact IDs selected by the agent.compose_stackvalidates and records that selection; missing or cautionary metadata must never make a canonical skill unselectable or unusable. - Keep the supported public path at manifest validation and immutable plan preview. Planning may write only the requested plan artifact; it must not materialize skill payloads or AAS managed state in the target.
- Treat apply and recovery as experimental opt-ins outside the supported preview claim. Do not add apply/recovery, benchmark, fuzz, crash/race, or synthetic verifier work unless the user explicitly places it in scope.
- When the task asks for end-to-end client proof, use a real supported client that discovers and invokes the local AAS MCP tools; direct stdio probes and automated tests do not substitute for that evidence.
- Do not tag, publish npm, deploy Pages, or write real user MCP configuration without the separately required publication approval.
Protected Release
Release only when requested.
- Include the target changelog entry in the maintainer batch PR so it is already on protected
main; avoid a separate release-notes-only PR. - From clean, current
main, runnpm run release:preflightand required security checks. - Run
npm run release:prepare -- X.Y.Z. This creates and pushesrelease/vX.Y.Zand opens the protected release PR. - Merge that release PR through its required checks, update local
mainto equalorigin/main, and wait for any canonical-sync PR to close. - Run
npm run release:publish -- X.Y.Z. It verifies the exact protected merge before creating or reusing the tag and GitHub Release. - Wait for publishing workflows, then verify the tag/ref, GitHub Release, npm version and dist-tag, CI, Pages, CodeQL, live
llms.txt,skills.json, and changed catalog routes.
Never rebase a published release tag, force stale release state, reuse a failed published version, or claim npm publication from the GitHub Release alone.
Stop Condition
Finish only when:
- every in-scope PR, issue, and alert is resolved or has one exact blocker;
- no open source or canonical-sync PR remains unintentionally;
main,origin/main, required workflows, generated state, and public surfaces agree;- the source and legacy repositories have no unintended infrastructure PR, their protected branches and Actions settings remain enforced, and the live manifest identifies the source repository;
- the user worktree is unchanged except for files the user explicitly placed in scope;
- release proof is complete when a release was requested.
Failure Rules
- A protected-branch rejection means switch to the PR path; never retry direct
mainpushes. - A missing PR checklist is informational; never mutate, close, or reopen a PR merely to refresh template metadata.
- Preserve unrelated dirty files and never stage them into maintainer work.
- Do not bypass
merge:batch, canonical-sync, or scripted release commands with generic Git helpers. - Do not weaken a test or policy gate merely to make a batch pass. Retire a gate only after explicit maintainer authorization, then update branch protection, workflow files, merge automation, documentation, and maintainer skills together so no phantom requirement remains.
Examples
For a reviewed source PR whose exact head is 0123456789abcdef0123456789abcdef01234567, exercise the protected path before merging:
npm run merge:batch -- --prs 914 --dry-run --reviewed-head 0123456789abcdef0123456789abcdef01234567
Run the same command without --dry-run only after every required check passes and the attested head remains unchanged.
Limitations
- This skill orchestrates the repository's existing scripts and protected workflows; it does not grant GitHub, npm, Pages, or local-client permissions.
- Stop at the exact approval or credential boundary when publication, authenticated configuration, or another externally visible action was not authorized.
- Re-read the current repository policy and
package.jsonon every run because branch protection, checks, and supported preview commands may change.